WO2018203568A1 - Cell evaluation system and method, and cell evaluation program - Google Patents

Cell evaluation system and method, and cell evaluation program Download PDF

Info

Publication number
WO2018203568A1
WO2018203568A1 PCT/JP2018/017499 JP2018017499W WO2018203568A1 WO 2018203568 A1 WO2018203568 A1 WO 2018203568A1 JP 2018017499 W JP2018017499 W JP 2018017499W WO 2018203568 A1 WO2018203568 A1 WO 2018203568A1
Authority
WO
WIPO (PCT)
Prior art keywords
cell
measurement information
information
cells
association
Prior art date
Application number
PCT/JP2018/017499
Other languages
French (fr)
Japanese (ja)
Inventor
昌士 鵜川
踊子 板橋
禎生 太田
Original Assignee
シンクサイト株式会社
国立大学法人 東京大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by シンクサイト株式会社, 国立大学法人 東京大学 filed Critical シンクサイト株式会社
Priority to US16/610,481 priority Critical patent/US11598712B2/en
Priority to JP2019515746A priority patent/JP7176697B2/en
Publication of WO2018203568A1 publication Critical patent/WO2018203568A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
    • G01N33/483—Physical analysis of biological material
    • G01N33/4833—Physical analysis of biological material of solid biological material, e.g. tissue samples, cell cultures
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume, or surface-area of porous materials
    • G01N15/10—Investigating individual particles
    • G01N15/14—Electro-optical investigation, e.g. flow cytometers
    • G01N15/1429—Electro-optical investigation, e.g. flow cytometers using an analyser being characterised by its signal processing
    • C—CHEMISTRY; METALLURGY
    • C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
    • C12M—APPARATUS FOR ENZYMOLOGY OR MICROBIOLOGY; APPARATUS FOR CULTURING MICROORGANISMS FOR PRODUCING BIOMASS, FOR GROWING CELLS OR FOR OBTAINING FERMENTATION OR METABOLIC PRODUCTS, i.e. BIOREACTORS OR FERMENTERS
    • C12M1/00—Apparatus for enzymology or microbiology
    • C12M1/34—Measuring or testing with condition measuring or sensing means, e.g. colony counters
    • G01N15/1433—
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume, or surface-area of porous materials
    • G01N15/10—Investigating individual particles
    • G01N15/14—Electro-optical investigation, e.g. flow cytometers
    • G01N15/1468—Electro-optical investigation, e.g. flow cytometers with spatial resolution of the texture or inner structure of the particle
    • G01N15/147—Electro-optical investigation, e.g. flow cytometers with spatial resolution of the texture or inner structure of the particle the analysis being performed on a sample stream
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
    • G01N33/483—Physical analysis of biological material
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G16B40/10—Signal processing, e.g. from mass spectrometry [MS] or from PCR
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume, or surface-area of porous materials
    • G01N15/10—Investigating individual particles
    • G01N2015/1006—Investigating individual particles for cytology
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G16B40/20—Supervised data analysis
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G16B40/30—Unsupervised data analysis

Definitions

  • the present invention relates to a cell evaluation system and method suitable for physically measuring one or more cells from a cell group and automatically evaluating the cells, and a cell evaluation program.
  • the smallest unit constituting a living organism is a cell.
  • a flow cytometry method has been proposed as a cell measurement technique in the case of performing analysis using such a single cell (single cell).
  • This flow cytometry method is a technique in which individual cells are dispersed in a fluid, and the fluid is finely flowed down and optically analyzed.
  • An apparatus using this technique is called a flow cytometer.
  • this flow cytometry method by irradiating excitation light while flowing fine particles such as cells to be observed in the flow path at high speed, and obtaining the total amount of fluorescence brightness and scattered light emitted from individual cells, The observation object can be evaluated.
  • Patent Document 1 discloses a flow cytometer and a flow cytometry method using the flow cytometer.
  • Fluorescence microscopes and imaging cytometers are also known as methods for observing individual cell phenotypes in more detail. These observation methods can acquire not only one-dimensional information such as the fluorescence luminance of the observation object and the total amount of scattered light but also two-dimensional and three-dimensional form information.
  • the observation object since the observation object does not move, it is difficult to perform a large amount of single cell analysis with high throughput.
  • an imaging flow cytometer capable of photographing cell morphological information at high speed with a throughput equivalent to that of a conventional flow cytometer (see, for example, Patent Document 2).
  • Cell phenotypes can be evaluated with spatial information of dimensions. This makes it possible to dramatically improve the amount of information in cell analysis while maintaining the throughput of the existing flow cytometry method, and to improve the quality and amount of cell phenotype analysis.
  • Patent Document 3 a method for evaluating and classifying cells using machine learning for the cell shape information has also been proposed (for example, Patent Document 3). reference.). Specifically, a corrector information (teacher data) is given in advance and a classifier is created. Then, supervised machine learning that evaluates and classifies the given cell information, or there is no correct answer information in advance. There is a method using unsupervised machine learning for evaluating and classifying cell information.
  • the present invention has been devised in view of the above-described problems, and the object of the present invention is to physically measure one or more cells from a group of cells and evaluate the evaluation in biology. It is an object to provide a cell evaluation system and method, and a cell evaluation program capable of automatically evaluating and classifying cells at high speed, accurately and simply by correctly attaching the correct measurement information.
  • the present inventors have three or more steps of reference measurement information that is stored in advance in a database and that is associated with biological measurement information for evaluating cells.
  • the reference measurement information is searched based on the newly measured measurement information of the cell with reference to the association degree of the cell, and the cell is measured by the biological measurement information linked to the searched measurement information for reference.
  • the cell evaluation system is a physical evaluation system that physically measures one or more cells in the cell evaluation system that physically measures one or more cells from a cell group and evaluates the cells.
  • the degree of association of three or more levels of measurement means, measurement information measured by the physical measurement means, and reference measurement information associated with biological measurement information for evaluating cells is stored in advance. And searching for the reference measurement information based on the cell measurement information newly measured through the physical measurement means with reference to the stored database and the association degree stored in the database. And an evaluation means for evaluating the cells based on the biological measurement information linked to the reference measurement information.
  • the cell evaluation system is a cell evaluation system that physically measures one or more cells from a group of cells and evaluates the cells, and physically measures the one or more cells.
  • a database in which three or more levels of association between measurement means, measurement information measured by the physical measurement means, and biological measurement information for evaluating cells are stored in advance, and stored in the database. And searching for the biological measurement information based on the measurement information of the cell newly measured through the physical measurement means with reference to the association degree, and the cell based on the searched biological measurement information. And an evaluation means for evaluating.
  • the cell evaluation system to which the present invention is applied is a cell evaluation system that physically measures one or more cells from a cell group and evaluates the cells.
  • structured excitation light or illumination light is irradiated, and the individual cells interact with the excitation light or illumination light in time series, and the optical spatial information of the cells is mapped to time series waveforms.
  • a physical measurement unit that physically measures, a database in which three or more levels of association between the measurement information measured by the physical measurement unit and the cell identification information are stored in advance, and the database
  • An evaluation means for referring to the stored association degree and evaluating by specifying the cell identification information based on the measurement information of the cell newly measured through the physical measurement means.
  • the cell evaluation method according to the present invention is a cell evaluation method in which one or more cells are physically measured from a cell group and the cells are evaluated. Newly measured cells are stored in advance in the database with three or more degrees of association with the reference measurement information associated with the biological measurement information, and the association degree stored in the database is referred to. The reference measurement information is searched based on the measurement information, and the cell is evaluated by biological measurement information linked to the searched reference measurement information.
  • the cell evaluation method is a cell evaluation method in which one or more cells are physically measured from a cell group and the cells are evaluated. Is measured physically by irradiating the excited excitation light or illumination light, interacting with the excitation light in time series for each cell, and mapping the optical spatial information of the cell to the time series waveform. Based on the measurement information of the newly measured cell, the degree of association of three or more levels of the measured measurement information and the cell identification information is stored in the database in advance and the association degree stored in the database is referred to. Thus, evaluation is performed by specifying cell specifying information.
  • the cell evaluation program according to the present invention is a cell evaluation program for physically measuring one or more cells from a cell group and evaluating the cells. And the reference measurement information associated with the measurement information for each reference linked to the biological measurement information, and searching for the reference measurement information based on the newly measured cell measurement information, The computer is caused to evaluate the cell based on the biological measurement information linked to the searched measurement information for reference.
  • the cell evaluation program according to the present invention is a cell evaluation program that physically measures one or more cells from a cell group and evaluates the cells. Is measured physically by irradiating the excited excitation light or illumination light, interacting with the excitation light in time series for each cell, and mapping the optical spatial information of the cell to the time series waveform. Based on the measurement information of the newly measured cell, the degree of association of three or more levels of the measured measurement information and the cell identification information is stored in the database in advance and the association degree stored in the database is referred to. Thus, the evaluation is performed by specifying the cell specifying information and causing the computer to execute the evaluation.
  • reference measurement information is selected by referring to the association degree described above from cell measurement information newly acquired via the imaging flow cytometer analysis unit. It becomes possible to discriminate the cell type from the reference measurement information via the biological measurement information and to evaluate the cell.
  • FIG. 1 shows a first embodiment of a cell evaluation system 1 to which the present invention is applied.
  • the cell evaluation system 1 includes a light source 2, an imaging flow cytometer analysis unit 3 that is irradiated with light from the light source 2, a light receiving unit 4 that receives optical information from the imaging flow cytometer analysis unit 3, A control unit 5 connected to the imaging flow cytometer analysis unit 3 and the light receiving unit 4, an evaluation unit 6 connected to the light source 2, the light receiving unit 4 and the control unit 5, and an evaluation unit 6 and a light receiving unit 4, respectively.
  • the database 7 is provided.
  • the light source 2 emits light necessary for analysis in the imaging flow cytometer analysis unit 3. Based on the imaging flow cytometry method, the imaging flow cytometer analysis unit 3 disperses individual cells in a fluid, finely flows down the fluid, and optically analyzes the cells to obtain cell shape information. Details of the imaging flow cytometer analysis unit 3 will be described later.
  • the light receiving unit 4 includes a sensor for receiving optical information obtained through the imaging flow cytometry method performed by the imaging flow cytometer analyzing unit 3.
  • the control unit 5 serves as a central control unit for controlling the light source 2, the imaging flow cytometer analysis unit 3, the light receiving unit 4, and the evaluation unit 6.
  • This control part 5 is comprised with a personal computer (PC), a portable terminal, a smart phone, a wearable terminal, a tablet type terminal etc., for example.
  • the evaluation unit 6 acquires the optical information obtained from the light receiving unit 4, and further refers to the information stored in the database to evaluate the cells flowing down in the imaging flow cytometer.
  • This cell evaluation includes cell type discrimination, cell function and characteristics, and the like.
  • the evaluation unit 6 includes a PC, a mobile terminal, a smartphone, a wearable terminal, a tablet terminal, and the like.
  • this evaluation part 6 may be comprised with the same device as the control part 5 mentioned above.
  • the database 7 is composed of a hard disk or the like for storing information necessary for evaluating cells by the evaluation unit 6 described above.
  • FIG. 2 shows a detailed configuration of the imaging flow cytometer analysis unit 3.
  • the imaging flow cytometer analysis unit 3 acquires a three-dimensional fluorescence image for each cell 8 flowing in the flow path 31.
  • the imaging flow cytometer analysis unit 3 includes a first objective lens 32, a first beam splitter 33, a first lens 34, a second lens 35, a second beam splitter 36, a second objective lens 38, A mirror 39, a third lens 40, and a fourth lens 41 are provided.
  • the first objective lens 32 focuses the light supplied from the light source 2 into the flow path 31, collects the return light reflected from the flow path 31, and sends this to the first beam splitter 33. .
  • the first beam splitter 33 plays a role of allowing the light emitted from the light source 2 to pass through as it is and reflecting the return light from the first objective lens 32.
  • the first lens 34 and the second lens 35 adjust the focal position and spot diameter of the return light reflected from the first beam splitter 33.
  • the second beam splitter 36 reflects a part of the return light that has passed through the second lens 35 toward the second objective lens 38 and passes a part of the return light as it is.
  • the second objective lens 38 guides the return light reflected from the second beam splitter 36 to the mirror 39, collects the return light reflected from the mirror 39, and passes this through the second beam splitter 36. Guide to the fourth lens 41.
  • the mirror 39 reflects the return light from the second objective lens 38.
  • the third lens 40 condenses the return light that has passed through the second beam splitter 36 and forms an image on the first light receiving sensor 4 a that constitutes the light receiving unit 4.
  • the fourth lens 41 condenses the return light that has passed through the second beam splitter 36 and forms an image on the second light receiving sensor 4 b that constitutes the light receiving unit 4.
  • Each component of the imaging flow cytometer analysis unit 3 operates based on control by the control unit 5.
  • the imaging flow cytometer analysis unit 3 may be one that uses the disclosed technology of US2015 / 0192767A1.
  • the first light receiving sensor 4a captures an xy plane image of the cell 8 based on the optical information of the formed return light.
  • the second light receiving sensor 4b captures an xz plane image of the cell 8 based on the optical information of the formed return light.
  • the first light receiving sensor 4a and the second light receiving sensor 4b convert the received light information of each planar image into an electric signal and transmit it to the evaluation unit 6 and the database 7.
  • the first light receiving sensor 4a and the second light receiving sensor 4b operate based on control by the control unit 5, respectively.
  • the imaging flow cytometer analysis unit 3 physically measures one or more cells flowing in the flow channel 31 based on the above-described configuration.
  • the physical measurement of the cell 8 includes visible image, electromagnetic wave, fluorescence, phase, transmission, spectroscopy, multicolor, scattering, reflection, coherent Raman, Raman or absorption / scattering / transmission / fluorescence spectrum, sound, terahertz, impedance Through either.
  • a fluorescence image is measured three-dimensionally will be described as an example.
  • the type and form of the cell 8 to be measured are not particularly limited as long as the effects of the present invention are not hindered, and the cell to be measured can be selected according to the purpose. Therefore, the cell 8 as the measurement target of the present invention may be a floating cell or an adherent cell. Bright cells also include bacteria (single cell organisms). Two or more cells may be measured, and in such a case, a mass (spheroid or the like) in which a plurality of cells are collected may be measured.
  • compartments 91 including at least one cell 8 and at least one bead 9 are prepared, and the three-dimensional cell 8 in each compartment 91 is prepared. It may be possible to repeat measurement of both the imaging information of the fluorescent image and the imaging information of the beads 9 over time.
  • the imaging information of the beads 9 in each compartment 91 is used as an index for specifying the cell 8 in each compartment 91.
  • Visible light from the light source 2 is irradiated from the first objective lens 32 to the cells 8 flowing through the flow path 31, and the return light is collected by the first objective lens 32, and passes through the first beam splitter 33. Reflected and passes through the first lenses 34 and 35. Further, a part of the return light passes through the second beam splitter 36 as it is, and forms an image on the first light receiving sensor 4 a via the third lens 40. Further, another part of the return light is reflected by the second beam splitter 36, reflected by the mirror 39, and imaged on the second light receiving sensor 4b via the fourth lens 41.
  • an xy plane image of the cell 8 is captured by the first light receiving sensor 4a, and an xz plane image of the cell 8 is captured by the second light receiving sensor 4b.
  • the three-dimensional fluorescence of the cell 8 is captured. An image will be formed.
  • the imaging information is continuously performed to obtain optical information of the fluorescence image of the three-dimensional whole image of the cell 8. It is also possible to do.
  • the measurement information obtained through the first light receiving sensor 4 a and the second light receiving sensor 4 b in this way is sent to the evaluation unit 6 and the database 7.
  • the evaluation unit 6 newly evaluates the cell 8 by analyzing physical measurement information of the cell 8 sent from the first light receiving sensor 4a and the second light receiving sensor 4b.
  • the evaluation unit 6 refers to information stored in the database 7 in the course of this evaluation.
  • the database 7 stores in advance three or more levels of association between measurement information and reference measurement information.
  • the cell images as measurement information imaged by the imaging flow cytometer analysis unit 3 described above are arranged on the left side through this association degree, and the reference measurement information is arranged on the right side through this association degree.
  • the reference measurement information is a list of typical examples of so-called cell images. This reference measurement information is classified into, for example, images AK according to the morphological characteristics of the cells.
  • the degree of association is determined based on whether a cell image as measurement information newly imaged through the imaging flow cytometer analysis unit 3 or the like is used as any of the images A to K previously defined for each cell form as the reference measurement information. It indicates whether the relevance is high. Further, the cell image as the measurement information is also categorized in advance into images P11 to P16 as shown in FIG. In this way, the measurement information and the reference measurement information are classified in advance for each image, and the classified measurement information images P11 to P16 and the reference measurement information images A to K are in three stages. They are associated with each other through the association described above. In other words, the degree of association is an index indicating whether there is a high possibility that the image of the measurement information is associated with the image of the reference measurement information. In selecting the reference measurement information from the measurement information, It shows the accuracy.
  • the measurement information image P11 has an association degree of 80% with the reference measurement information image A, an association degree with the image B of 60%, an association degree with the image H of 40%, and an association degree with the image J of 20%.
  • the measurement information image P13 has an association degree of 100% with the reference measurement information image D, an association degree of the image E of 80%, and an association degree of the image F of 60%.
  • the measurement information image P15 has an association degree of 80% with the reference measurement information image H, an association degree with the image D of 40%, and an association degree with the image A of 20%.
  • the thickness of the line connecting between the measurement information and the reference measurement information in FIG. 4 indicates the magnitude of the degree of association, and the line is not connected between the measurement information and the reference measurement information. In this case, it means that the degree of association is 0%.
  • association degree may be constituted by a model that can be updated through so-called machine learning, or may be constituted by a neural network. In addition, this association degree may be configured by a network on the assumption that deep learning is performed.
  • the evaluation unit 6 refers to the association degree stored in the database 7 in this manner, and the measurement information of the cell 8 newly acquired via the light receiving unit 4 (first light receiving sensor 4a, second light receiving sensor 4b) is obtained. It is determined which reference measurement information is likely to correspond. If the measurement information of the cell 8 newly acquired via the light receiving unit 4 is similar to the image P14 accumulated in advance in the measurement information, this image P14 and the reference measurement information image G having a high degree of association are most applicable. It is determined that there is a high possibility of the image C, and next, it is determined that there is a high possibility that the image C is applicable.
  • this image P16 and the reference measurement information image J having a high degree of association are most applicable. It is determined that there is a high possibility of the image K, and next, it is determined that there is a high possibility of corresponding to the image K.
  • the evaluation unit 6 performs an operation of selecting an image of the reference measurement information by referring to these association degrees based on the measurement information of the cell 8 newly acquired through the light receiving unit 4. At this time, the evaluation unit 6 may select an image of reference measurement information having the highest degree of association with the measurement information of the cell 8 newly acquired via the light receiving unit 4. This is because, as described above, the higher the degree of association, the higher the accuracy of the selection. However, the evaluation unit 6 is not limited to the case of selecting the reference measurement information image having the highest degree of association. The evaluation unit 6 may intentionally select one having a medium degree of association or a low degree of association. May be.
  • reference measurement information having an association degree of 0% in which no arrow is connected between the measurement information and the reference measurement information may be selected.
  • the evaluation unit 6 is not limited to the case where one piece of reference measurement information is selected, and a plurality of pieces of reference measurement information may be selected after referring to the association degree.
  • the evaluation unit 6 identifies the cell type from the selected reference measurement information.
  • FIG. 5 shows a process until the cell type is specified from the selected reference measurement information.
  • Each reference measurement information is associated with biological measurement information.
  • the reference measurement information and the biological measurement information are usually linked one-to-one. However, the reference measurement information and the biological measurement information are not limited to this. One-to-multiple, multiple-to-one, or multiple-to-multiple are linked to each other. It may be done.
  • This biological measurement information is information measured via, for example, transcriptome, genome, epigenome, protein, metabolite, sugar, lipid, time-series change tendency of cells, etc. in order to specify the cell type. It is.
  • This biological measurement information is information acquired in advance for each cell of the images AK constituting the reference measurement information via a transcriptome, a genome, or the like. By acquiring this biological measurement information for each cell in the images AK, it is possible to grasp the exact cell type of each cell in the images AK.
  • Each biological measurement information and the cell type identified therefrom usually correspond with 1: 1. In the example of FIG. 5, when the biological measurement information of the cell of the image A for reference measurement information is “genome OO”, the cell type a is determined from the “genome OO”.
  • this biological measurement information includes optical imaging measurement information for acquiring finer or different optical information. That is, biological measurement information is a concept that includes imaging (morphological) analysis results in addition to cell types and the like.
  • the database 7 stores biological measurement information corresponding to each reference measurement information as shown in FIG. 5, and further stores cell types specified from the biological measurement information. Therefore, the evaluation unit 6 refers to the database 7 to determine the biological measurement information associated with the selected reference measurement information, and selects the cell type specified by the determined biological measurement information. Determine.
  • the evaluation unit 6 can select the reference measurement information as shown in FIG. 6, the cell type can be immediately discriminated through this biological measurement information.
  • the evaluation unit 6 displays the determined cell type via a display unit such as a display (not shown).
  • a display unit such as a display (not shown).
  • the user can immediately grasp the newly acquired cell type of the cell 8 by visually recognizing a display unit (not shown).
  • the reference measurement information is selected by referring to the association degree described above from the cell measurement information newly acquired via the imaging flow cytometer analysis unit 3.
  • the cell type can be discriminated from the reference measurement information via the biological measurement information.
  • these evaluation operations can be automatically performed without manual intervention. This makes it possible to quickly and accurately classify and evaluate cell information based on biologically correct correct information, based on physical measurement information, regardless of the observer's subjective judgment.
  • the cell evaluation system 1 to which the present invention is applied is characterized in that the reference measurement information is searched through the association degree set in three or more stages.
  • the association degree can be described by a numerical value of, for example, 0 to 100%, but is not limited to this, and may be configured at any stage as long as it can be described by a numerical value of three or more levels.
  • the present invention it is possible to make a determination without missing an extremely low reference measurement information having a relevance of 1%. Even if the measurement information for reference is extremely low, it is connected as a small sign, and it is useful to the user that it may be useful as measurement information for reference once every tens or hundreds of times. You can call attention.
  • a search policy can be determined in a manner of setting a threshold by performing a search based on such three or more levels of association. If the threshold value is lowered, it can be picked up without omission even if the association degree is 1%, but there are cases where many cell types are picked up based on the reference measurement information with a low possibility of correct answer. On the other hand, if the threshold value is increased, only the cell types specified by the reference measurement information that is highly likely to be correct can be narrowed down. On the other hand, the reference measurement that displays a suitable solution once every tens or hundreds of times. The cell type specified by the information may be overlooked. It is possible to decide which to place importance on the basis of the idea on the user side and the system side, but it is possible to increase the degree of freedom in selecting points to place such emphasis.
  • the measurement information in FIG. 3 has been described by taking as an example the case where the association degree of the reference measurement information is set in advance for each of the measurement information P11 to P16. It is not done. Three or more levels of association between the combination of measurement information and the reference measurement information may be stored in advance, and evaluation may be performed based on two or more measurement information newly physically measured.
  • the association degree of the reference measurement information D is defined as 60%
  • the association degree of the reference measurement information E is defined as 40%
  • the like the association degree of the reference measurement information
  • the relevance of the reference measurement information H is 20%
  • the relevance of the reference measurement information I is 50%
  • the relevance of the reference measurement information J is 70%. Etc. are defined.
  • the association degree of the reference measurement information D is 60%
  • the reference measurement information E It can be determined that the degree of association is 40% or the like.
  • the reference measurement information is similarly selected with reference to these degrees of association.
  • the association degree of the reference measurement information H is It can be determined that 20%
  • the association degree of the reference measurement information I is 50%
  • the association degree of the reference measurement information J is 70%.
  • the reference measurement information is similarly selected with reference to these degrees of association.
  • the example of referring to the three or more levels of association between the combination of measurement information and the reference measurement information is applied to so-called multi-instance learning or the like that performs comparison verification between a set of captured images of cells. Also good.
  • each cell image related to a certain disease is stored as reference measurement information.
  • a plurality of cell images as measurement information are captured for each of a plurality of cells of a patient A, and it is determined whether or not the disease corresponds to the reference measurement information by referring to the association degree. It may be.
  • the association degree described above may be updated. That is, measurement information as shown in FIG. 4 and reference measurement information are updated as needed.
  • This update may reflect information provided via a public communication network such as the Internet.
  • the system side or the user side may be updated artificially or automatically based on the contents of research data, papers, conference presentations, newspaper articles, books, etc. by experts. Artificial intelligence may be used in these update processes.
  • This renewal of the association increases or decreases the association every time information about the relationship between the measurement information and the reference measurement information is entered. For example, if it is newly confirmed that a certain measurement information image corresponds to a certain reference measurement information image through research data based on a paper, a conference presentation, or other experimental verification, the measurement information image And the degree of association between the reference measurement information image and the image of the reference measurement information is increased. In addition, if it is newly confirmed that a measurement information image does not correspond to a reference measurement information image through research papers based on papers, conference presentations, or other experimental verification, the measurement information image And the relevance of the reference measurement information image is lowered.
  • the degree of relevance is set to three or more levels, so that it is possible to freely deal with cases where it is desired to increase or decrease the degree of relevance.
  • This relevancy update itself may be performed through the above-described machine learning and deep learning.
  • the present invention is not limited to the example of specifying the cell type, and may be applied when evaluating the function and characteristics of the cell.
  • what is specified by the biological measurement information shown in FIG. 5 is not a cell type but a cell function or a cell characteristic.
  • the correspondence relationship of each cell function and each cell characteristic with respect to each biological measurement information is acquired in advance and stored in the database 7.
  • FIG. 6 when an image of reference measurement information is selected, each function or characteristic of the cell that is immediately replaced by the cell type is immediately determined through this biological measurement information. Will be determined.
  • the case where measurement information is acquired via a two-dimensional or three-dimensional fluorescence image has been described as an example, but other electromagnetic waves, bright field, dark field, fluorescence, phase, transmission, The same applies when cell 8 is physically measured by spectroscopy, multicolor, scattering, reflection, coherent Raman, Raman or absorption / scattering / transmission / fluorescence spectrum, sound, terahertz, impedance, and measurement information is obtained from it. It is.
  • the reference measurement information is also prepared in advance according to the physical measurement means, and is associated with each other with the degree of association.
  • both the measurement information and the reference measurement information are fluorescence spectra.
  • the peak position and the change tendency with respect to each wavelength are classified in advance.
  • the reference measurement information is selected by referring to the association degree based on the measurement information including the fluorescence spectrum newly acquired from the cell 8 for the categorized fluorescence spectrum pattern as the reference measurement information. It becomes.
  • the physical measurement means can be based on any method such as an observation method using a microscope including a compartment observation method in a microwell using a microscope, an imaging cytometry method, etc. Also good. Further, the present invention may be applied to the case of measuring cells that are not in the liquid in addition to the case of physically measuring the cells flowing down in the liquid.
  • the microscope used may be an optical microscope (including bright field, phase difference, fluorescence, confocal laser, Raman, etc.) or an electron microscope (transmission type, scanning type).
  • the degree of association of three or more levels of measurement information and biological measurement information may be stored in advance.
  • the image of the cell as measurement information imaged by the imaging flow cytometer analysis unit 3 described above is arranged on the left side via this association degree, and the biological measurement information is arranged on the right side via this association degree.
  • the biological measurement information includes, for example, various types of genome information for specifying cell types (cell functions, characteristics, etc.). This biological measurement information is categorized in advance according to genome information and the like, and genome information O to genome information Y and the like are classified as shown in FIG.
  • This degree of association is related to any of the genome information O to genome information Y, etc., that the cell image as measurement information newly imaged through the imaging flow cytometer analysis unit 3 etc. is previously classified as this reference measurement information Are related to each other through three or more levels of association.
  • the degree of association is an index indicating which biological measurement information is likely to be associated with an image of measurement information. In selecting biological measurement information from measurement information, It shows the accuracy.
  • the measurement information image P11 has an association degree of 80% with the genome information O of the biological measurement information, an association degree of 60% with the genome information P, an association degree of 40% with the genome information V, and an association degree of 20% with the genome information X. It is shown that.
  • the measurement information image P13 has an association degree of 100% with the genome information R of the biological measurement information, an association degree of 80% with the genome information S, and an association degree of 60% with the genome information T.
  • the thickness of the line connecting the measurement information and the biological measurement information in FIG. 7 indicates the magnitude of the association, and the line is connected between the measurement information and the reference measurement information. If not, it means that the relevance is 0%.
  • this association degree may also be constituted by a model that can be updated through so-called machine learning, or may be constituted by a neural network. In addition, this association degree may be configured by a network on the assumption that deep learning is performed.
  • the evaluation unit 6 refers to the association degree stored in the database 7 in this manner, and the measurement information of the cell 8 newly acquired via the light receiving unit 4 (first light receiving sensor 4a, second light receiving sensor 4b) is obtained. It is determined which biological measurement information is likely to correspond. If the measurement information of the cell 8 newly acquired via the light receiving unit 4 is similar to the image P14 accumulated in advance in the measurement information, the genome information U of the biological measurement information highly related to the image P14 is included. It is determined that there is a high possibility that it corresponds to the most, and next, it is determined that there is a high possibility that it corresponds to the genome information Q.
  • the genome information X of the biological measurement information highly related to the image P16 is used. It is determined that there is a high possibility that it corresponds to the most, and then it is determined that there is a high possibility that it corresponds to the genome information Y.
  • the evaluation unit 6 performs an operation of selecting biological measurement information by referring to these association degrees based on the measurement information of the cell 8 newly acquired via the light receiving unit 4. At this time, the evaluation unit 6 may select biological measurement information having the highest degree of association with the measurement information of the cell 8 newly acquired through the light receiving unit 4. This is because, as described above, the higher the degree of association, the higher the accuracy of the selection.
  • the evaluation unit 6 is not limited to selecting an image of biological measurement information with the highest degree of association, and it may be selected with a medium degree of association or a low degree of association. It may be. In addition to this, it is needless to say that biological measurement information having an association degree of 0% in which no arrow is connected between measurement information and biological measurement information may be selected.
  • the evaluation unit 6 is not limited to selecting one biological measurement information, but may select a plurality of reference measurement information after referring to the association degree.
  • the evaluation unit 6 identifies the cell type from the selected biological measurement information. Since the biological measurement information is originally associated with the cell type, if this biological measurement information can be selected, the cell type can be identified immediately. It is also possible to evaluate cells other than the cell type by associating in advance the evaluation of the function and characteristics of the cells other than the cell type with the biological measurement information.
  • Such association between the measurement information shown in FIG. 7 and the biological measurement information may be updated.
  • This update may reflect information provided via a public communication network such as the Internet.
  • the system side or the user side may be updated artificially or automatically based on the contents of research data, papers, conference presentations, newspaper articles, books, etc. by experts. Artificial intelligence may be used in these update processes.
  • This renewal of association increases or decreases the association every time information about the relationship between measurement information and biological measurement information is entered. For example, if an image of certain measurement information can be newly confirmed through research papers based on papers, conference presentations, or other experimental verifications, the measurement information corresponds to the genomic information of a certain biological measurement information. Increase the degree of association between the image and the genome information of biological measurement information. In addition, if it is newly confirmed that a measurement information image does not correspond to the genomic information of a certain biological measurement information through a research paper based on a paper, a conference presentation, or other experimental verification, the measurement information The degree of association between the image and the genome information of biological measurement information is lowered.
  • FIG. 8 shows a form of a cell evaluation system 1 ′ to which the present invention is applied.
  • the cell evaluation system 1 ′ includes an imaging flow cytometer analysis unit 3 ′, a lens 72 connected to the imaging flow cytometer analysis unit 3, and a PMT (one pixel element) 71 from which light is guided from the lens 72. Yes.
  • the cell evaluation system 1 ′ includes an evaluation unit 6 that evaluates data detected by the PMT 71, and a control unit 7 and a database 7 that are connected to the evaluation unit 6, respectively.
  • the known stationary random structure illumination unit 30 is an optical structure that performs so-called structured illumination that emits excitation light through a specific gray pattern.
  • This known stationary random structure illumination unit 30 is designed based on a design method using compressed sensing or any other design method.
  • the structured excitation light or illumination light can be irradiated through such an optical structure, and an optical pattern can be formed on the sample surface.
  • a cell group including the cells 8 to be measured flows down in the flow path in the known stationary random structure illumination unit 30. As a result, the excitation light is irradiated to the measurement target cell group flowing down in the flow path.
  • Each cell 8 passes through an optical element such as a diffraction element and interacts with excitation light in a time series.
  • Physical measurement can be realized by mapping the optical spatial information (information on the form) of the cell 8 to a time-series waveform.
  • the imaging flow cytometer analysis unit 3 ′ may apply a so-called ghost cytometry method. Further, the cell image may be reconstructed based on the obtained waveform signal, but the reconstruction of the cell image is not particularly essential in the second embodiment.
  • the excitation light excites different parts of the cell 8 in time series to cause interaction, and the spatial information of the cell 8 is compressed and converted into time information.
  • the movement of the cell 8 flowing in the imaging flow cytometer analysis unit 3 ′ and the emission of excitation light (illumination light) from the known stationary random structure illumination unit 30 proceed, and the excitation light (illumination light) overlaps with the cells.
  • the part that changes is changed, and the time change of the total amount of overlap is measured. That is, a time series waveform in which the horizontal axis is a time axis and the vertical axis is a signal intensity axis as shown in FIG. 8 can be detected.
  • Such a time series waveform is detected by the PMT 71, which is a single pixel element with high speed and high sensitivity.
  • the lens 72 performs signal measurement for one pixel by guiding the light from the imaging flow cytometer analysis unit 3 ′ to the PMT 71 and collecting it.
  • the evaluation unit 6 searches for cell specific information in the cells flowing through the imaging flow cytometer analysis unit 3 ′ based on the time-series waveform obtained in this way.
  • the cell specifying information here is a concept including any information for specifying a cell flowing through the imaging flow cytometer analysis unit 3 ′. For example, cell phenotype, cell characteristics, cell type, intracellular molecular localization, Information for defining intracellular structures, cell morphology, cell maturity, important cell characterization, and the like. In the following example, a case where the cell type of the cell flowing through the imaging flow cytometer analysis unit 3 ′ is searched as an example of the cell specifying information will be described.
  • the evaluator 6 constructs and stores a learned model in the database 7 in advance when searching for this cell identification information.
  • An example of this learned model is shown in FIG.
  • the time-series waveform is arranged on the left side through this association degree, and the cell identification information (cell type) is arranged on the right side through this association degree.
  • the degree of association is the same as in the example of FIG. 4 described above, but in this example, the time-series waveform indicates which of the cell identification information is more relevant.
  • U11 of the time-series waveform has an association degree of 80% with cell type A, an association degree of 60% with cell type B, an association degree of 40% with cell type H, and an association degree of 20% with cell type J. It has been shown.
  • U13 of the time-series waveform is shown to have an association degree of 100% with the cell type D of the cell identification information, an association degree of 80% with the cell type E, and an association degree of 60% with the cell type F.
  • U15 of the time-series waveform is shown to have an association degree of 80% with the cell type H of the cell identification information, an association degree of 40% with the cell type D, and an association degree of 20% with the cell type A.
  • the line is not connected between the time-series waveform and the cell specifying information, it means that the association degree is 0%.
  • association degree may be constituted by a model that can be updated through so-called machine learning, or may be constituted by a neural network. In addition, this association degree may be configured by a network on the assumption that deep learning is performed.
  • the evaluation unit 6 refers to the association degree stored in the database 7 in this manner, and determines which cell specifying information is likely to correspond to the time-series waveform of the cell 8 newly acquired via the PMT 71. Determine. If the newly acquired time-series waveform of the cell 8 is similar to the time-series waveform U14, it is determined that there is a high possibility that the cell type G of the cell identification information having a high degree of association with this U14 is the most relevant. Next, it is determined that there is a high possibility of corresponding to cell type C.
  • the time series waveform of the newly acquired cell 8 is similar to U16, it is determined that there is a high possibility that the cell type J of the cell identification information having a high degree of association with U16 is most likely, and then the cell It is determined that there is a high possibility that it corresponds to the seed K.
  • the evaluation unit 6 performs an operation of selecting any one or more of the cell identification information by referring to these association degrees based on the newly acquired time series information of the cells 8. At this time, the evaluation unit 6 may select cell specifying information having the highest degree of association with the time series information of the cell 8 newly acquired via the light receiving unit 4. This is because, as described above, the higher the degree of association, the higher the accuracy of the selection. However, the evaluation unit 6 is not limited to selecting the image of the cell identification information with the highest degree of association. The evaluation unit 6 may select the medium with the medium degree of association or the one with the low degree of association. Also good.
  • cell specifying information having an association degree of 0% in which no arrow is connected between the time series information and the cell specifying information may be selected.
  • the evaluation unit 6 is not limited to the case of selecting one piece of the cell specifying information, but may select a plurality of pieces of cell specifying information with reference to the association degree.
  • this degree of association may be replaced with a neuron of a neural network.
  • a learned model is constructed in which one or more cell identification information for a combination of a plurality of time-series waveforms is associated through association.
  • one or more pieces of cell specifying information may be selected based on the method described above.
  • FIG. 10 shows an example of learning by tagging time-series waveforms for each cell specifying information. That is, a time-series waveform is learned for each cell type A and B. A learned model in which a time series waveform is learned through three or more levels of association for each cell specifying information is stored. An example of such a learned model is the network shown in FIG. Next, it is determined as described above which cell identification information (cell type A, cell type B) the time series waveform of the cell 8 newly acquired via the PMT 71 corresponds to.
  • cell identification information cell type A, cell type B
  • FIG. 11 shows an example in which learning is performed through three or more levels of association between a time-series waveform and each cell specifying information about a mixture obtained by mixing a plurality of cell specifying information. That is, in the construction stage of the learned model, a time-series waveform for a mixture obtained by mixing the cell type A and the cell type B is acquired. The time-series waveform obtained as a result includes both a sign corresponding to the cell type A and a sign corresponding to the cell type B. Such a time series waveform is learned as a learned model, and a cell group in which cells 8 to be measured are combined is newly acquired via the PMT 71. From the time-series waveform obtained therefrom, a sign corresponding to the cell type A and a sign corresponding to the cell type B are respectively compared with the learned model to determine which corresponds.
  • the time series waveform is arranged on the left side of the association degree shown in FIG. 9, and this time series waveform is either one of the signs according to the cell type A and the signs according to the cell type B or Both signs are a combination.
  • cell specific information is tagged via the association degree to constitute a learned model.
  • FIG. 12 is an example of constructing a learned model in which a positive time series waveform and a negative time series waveform are learned in advance through three or more degrees of association in specifying cell identification information to be detected. Similar to the example of FIG. 11, in detecting specific cell specific information (for example, cell type B) from a time series waveform of a mixture obtained by mixing a plurality of cell specific information, a positive time series waveform and a negative A simple time-series waveform is learned in advance.
  • the positive time-series waveform here is a sign on the waveform having a high probability of being applied to specific cell specific information (for example, cell type B).
  • a negative time-series waveform is a sign on the waveform that has a low probability of being applied to specific cell specific information (for example, cell type B).
  • the time series waveform is arranged on the left side of the association degree shown in FIG. 9, and positive or negative is arranged on the right side of the association degree, and each is associated through the association degree.
  • a positive time series waveform sign appears as a result.
  • the probability of being applied to cell specific information for example, cell type B
  • the probability of being applied to specific cell specific information for example, cell type B
  • positive signs and negative signs that apply to this specific cell-specific information are comprehensively determined, and the probability of applying to the specific cell-specific information is determined.
  • a learned model in which only negative time-series waveforms are learned in advance through three or more levels of association may be stored.
  • a time series waveform is newly acquired from the cell group in which the cells 8 to be measured are combined via the PMT 71, it is determined with reference to the learned model whether the time series waveform corresponds to the negative time series waveform. .
  • information that does not correspond to a negative time-series waveform may be specified as cell specifying information to be detected.
  • FIG. 13 shows an example of creating a learned model by acquiring positive signs and negative signs together with cell identification information.
  • the cell specific information obtained through the green fluorescent live cell staining and the red fluorescent biomarker is simultaneously measured together with the time-series waveform exhibiting positive signs.
  • the cell specific information obtained through the green fluorescent live cell staining and the red fluorescent biomarker is simultaneously measured together with the time-series waveform exhibiting a negative sign.
  • time-series waveforms obtained in this way and the positive or negative cell specific information tagged to each are measured simultaneously and accumulated as a learned model.
  • the cell specifying information may be specified based on the learning. For example, it is useful when most of the labeling is known, such as cancer cells in blood, when most do not know how to label.

Abstract

[Problem] To physically measure one or more cells from a cell group and perform automatic evaluation thereof. [Solution] The present invention is characterized in that: the degree of relation of three or more stages with reference measurement information that is respectively associated with biological measurement information for evaluating a cell is referenced, said reference measurement information being stored in a database 7 in advance; the reference measurement information is retrieved on the basis of the measurement information of a newly measured cell; and the cell is evaluated using the biological measurement information associated with the retrieved reference measurement information.

Description

細胞評価システム及び方法、細胞評価プログラムCell evaluation system and method, cell evaluation program
　細胞群の中から一以上の細胞を物理的に計測してその細胞の評価を自動的に行う上で好適な細胞評価システム及び方法、細胞評価プログラムに関するものである。 The present invention relates to a cell evaluation system and method suitable for physically measuring one or more cells from a cell group and automatically evaluating the cells, and a cell evaluation program.
　生物を構成する最小単位は細胞である。従来より、生物の機能や構造、形態等について各種解明を行う場合には、あくまで細胞群を対象として行われ、これを構成する個々の細胞単位では行われてこなかった。 The smallest unit constituting a living organism is a cell. Conventionally, when various clarifications are made on the function, structure, form, etc. of a living organism, it has been carried out only on a cell group, and has not been carried out on an individual cell unit basis.
　しかし近年において、癌組織等において表現型が類似した細胞タイプにおいても、細胞毎に遺伝子発現が異なることが明らかになってきた。このため、細胞群単位で各種解明を行うのではなく、個々の細胞単位で解析を行う必要がある。 However, in recent years, it has become clear that gene expression differs from cell to cell even in cell types with similar phenotypes in cancer tissues and the like. For this reason, it is necessary to carry out analysis in units of individual cells rather than performing various clarifications in units of cell groups.
　このような一細胞（シングルセル）による解析を行う場合における細胞計測技術として、フローサイトメトリー法が提案されている。このフローサイトメトリー法は、個々の細胞を流体中に分散させ、その流体を微細に流下させて光学的に分析する技術であり、この技術を用いた装置をフローサイトメータと呼ぶ。このフローサイトメトリー法では、流路中に観察対象となる細胞等の微粒子を高速に流下しながら励起光を照射し、個々の細胞から発せられる蛍光輝度や散乱光の総量を取得することよって、観察対象物を評価することができる。このフローサイトメトリー法を用いると、非常に高いスループットで単一細胞の解析を行うことが可能となる。特許文献１には，フローサイトメータ及びそのフローサイトメータを用いたフローサイトメトリー法が開示されている。 A flow cytometry method has been proposed as a cell measurement technique in the case of performing analysis using such a single cell (single cell). This flow cytometry method is a technique in which individual cells are dispersed in a fluid, and the fluid is finely flowed down and optically analyzed. An apparatus using this technique is called a flow cytometer. In this flow cytometry method, by irradiating excitation light while flowing fine particles such as cells to be observed in the flow path at high speed, and obtaining the total amount of fluorescence brightness and scattered light emitted from individual cells, The observation object can be evaluated. When this flow cytometry method is used, it becomes possible to analyze a single cell with very high throughput. Patent Document 1 discloses a flow cytometer and a flow cytometry method using the flow cytometer.
　また、より詳細に個々の細胞の表現型を観察する方法として、蛍光顕微鏡やイメージングサイトメーターが知られている。これらの観察法は、観察対象物の蛍光輝度や散乱光の総量という１次元の情報だけではなく、２次元・３次元の形態情報まで取得できる。一方、フローサイトメータと異なり、観察対象物が移動しないために、ハイスループットに大量の単一細胞解析を行うことは難しい。この点、従来のフローサイトメータと同等のスループットで細胞の形態情報を高速に撮影できるイメージングフローサイトメーターが知られており（例えば、特許文献２参照。）、蛍光細胞画像を含む２次元・３次元の空間情報で細胞の表現型を評価することができる。これにより、既存のフローサイトメトリー法のスループットを維持しながら、細胞解析の情報量を飛躍的に向上させることが可能となり、細胞の表現型解析の質と量を向上させることができる。 Fluorescence microscopes and imaging cytometers are also known as methods for observing individual cell phenotypes in more detail. These observation methods can acquire not only one-dimensional information such as the fluorescence luminance of the observation object and the total amount of scattered light but also two-dimensional and three-dimensional form information. On the other hand, unlike the flow cytometer, since the observation object does not move, it is difficult to perform a large amount of single cell analysis with high throughput. In this regard, there is known an imaging flow cytometer capable of photographing cell morphological information at high speed with a throughput equivalent to that of a conventional flow cytometer (see, for example, Patent Document 2). Cell phenotypes can be evaluated with spatial information of dimensions. This makes it possible to dramatically improve the amount of information in cell analysis while maintaining the throughput of the existing flow cytometry method, and to improve the quality and amount of cell phenotype analysis.
　更に、イメージングフローサイトメーターによって生成される膨大な細胞形態情報を有効に活用するために、細胞形態情報に機械学習を用いて細胞を評価、分類する方法も提案されている（例えば、特許文献３参照。）。具体的には、事前に正解情報（教師データ）を与えて分類器を作成した上で、与えられた細胞情報の評価、分類を行う教師あり機械学習や、事前に正解情報がなく、与えられた細胞情報を評価、分類する教師なし機械学習等を用いる方法がある。 Furthermore, in order to effectively utilize the enormous amount of cell shape information generated by the imaging flow cytometer, a method for evaluating and classifying cells using machine learning for the cell shape information has also been proposed (for example, Patent Document 3). reference.). Specifically, a corrector information (teacher data) is given in advance and a classifier is created. Then, supervised machine learning that evaluates and classifies the given cell information, or there is no correct answer information in advance. There is a method using unsupervised machine learning for evaluating and classifying cell information.
特許第５５３４２１４号公報Japanese Patent No. 5534214 米国特許第６２４９３４１号公報US Pat. No. 6,249,341 特願２０１５－２１２３５６号Japanese Patent Application No. 2015-212356
　しかしながら、教師あり機械学習を用いる場合には、正解付けをする教師は人間であり、人間が主観的に細胞を評価、分類した結果を教師データとしていた。そのため、細胞の形態情報を含めた物理的計測手段によって取得された情報は、人間が差異を判別できる範囲でしか分類、評価できず、各個人が持つバイアスを排除できないという問題があった。また、教師なし機械学習では、人間の主観的判断によらず自動で分類、評価が可能であるものの、一方で機械がどのような判断基準で分類、評価しているかを説明することができず、生物学的に意味のある分類、評価結果なのかを説明することができなかった。 However, in the case of using supervised machine learning, the correct teacher is a human, and the result of human evaluation and classification of cells subjectively is used as teacher data. For this reason, there is a problem that information acquired by physical measurement means including cell morphology information can be classified and evaluated only within a range in which a person can discriminate a difference, and a bias of each individual cannot be eliminated. In addition, although unsupervised machine learning can automatically classify and evaluate regardless of human subjective judgment, it cannot explain what criteria the machine classifies and evaluates. I was unable to explain the biologically meaningful classification and evaluation results.
　また、従来の生物学的計測手法である次世代DNA・RNAシークエンシング、ポリメラーゼ連鎖反応（PCR）、DNAマイクロアレイ等を用いる場合には、細胞の核酸、タンパク質、代謝等の生化学的、分子生物学的情報に基づいているため、正確に細胞の分類、評価を行うことができる。しかし、これらの生物学的計測手法は、物理的計測手法に比べると遥かに高価な上に、スループットが遅く、膨大な量の単一細胞解析を行うことができないという問題点が生じる。 In addition, when using conventional biological measurement techniques such as next-generation DNA / RNA sequencing, polymerase chain reaction (PCR), DNA microarrays, etc., biochemical and molecular organisms such as cell nucleic acids, proteins, metabolism, etc. Since it is based on scientific information, it is possible to accurately classify and evaluate cells. However, these biological measurement methods are far more expensive than physical measurement methods, and have a problem that throughput is slow and a large amount of single cell analysis cannot be performed.
　そこで、本発明は、上述した問題点に鑑みて案出されたものであり、その目的とするところは、細胞群の中から一以上の細胞を物理的に計測して、その評価を生物学的計測情報によって正解付けすることで、高速、正確、かつ簡便に細胞の評価、分類を自動的に行うことが可能な細胞評価システム及び方法、細胞評価プログラムを提供することにある。 Therefore, the present invention has been devised in view of the above-described problems, and the object of the present invention is to physically measure one or more cells from a group of cells and evaluate the evaluation in biology. It is an object to provide a cell evaluation system and method, and a cell evaluation program capable of automatically evaluating and classifying cells at high speed, accurately and simply by correctly attaching the correct measurement information.
　本発明者らは、上述した課題を解決するために、データベースに予め記憶されている、細胞を評価するための生物学的計測情報にそれぞれが紐付けられた参照用計測情報との３段階以上の連関度を参照し、新たに計測された細胞の計測情報に基づいて、上記参照用計測情報を探索するとともに、探索した上記参照用計測情報に紐付けられた生物学的計測情報により上記細胞を評価する細胞評価システム及び方法、細胞評価プログラムを発明した。 In order to solve the above-mentioned problems, the present inventors have three or more steps of reference measurement information that is stored in advance in a database and that is associated with biological measurement information for evaluating cells. The reference measurement information is searched based on the newly measured measurement information of the cell with reference to the association degree of the cell, and the cell is measured by the biological measurement information linked to the searched measurement information for reference. Have invented a cell evaluation system and method, and a cell evaluation program.
　即ち、本発明に係る細胞評価システムは、細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価システムにおいて、上記一以上の細胞を物理的に計測する物理的計測手段と、上記物理的計測手段により計測された計測情報と、細胞を評価するための生物学的計測情報にそれぞれが紐付けられた参照用計測情報との３段階以上の連関度が予め記憶されているデータベースと、上記データベースに記憶されている連関度を参照し、新たに物理的計測手段を介して計測された細胞の計測情報に基づいて、上記参照用計測情報を探索するとともに、探索した上記参照用計測情報に紐付けられた生物学的計測情報により上記細胞を評価する評価手段とを備えることを特徴とする。 That is, the cell evaluation system according to the present invention is a physical evaluation system that physically measures one or more cells in the cell evaluation system that physically measures one or more cells from a cell group and evaluates the cells. The degree of association of three or more levels of measurement means, measurement information measured by the physical measurement means, and reference measurement information associated with biological measurement information for evaluating cells is stored in advance. And searching for the reference measurement information based on the cell measurement information newly measured through the physical measurement means with reference to the stored database and the association degree stored in the database. And an evaluation means for evaluating the cells based on the biological measurement information linked to the reference measurement information.
　また、本発明に係る細胞評価システムは、細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価システムにおいて、上記一以上の細胞を物理的に計測する物理的計測手段と、上記物理的計測手段により計測された計測情報と、細胞を評価するための生物学的計測情報との３段階以上の連関度が予め記憶されているデータベースと、上記データベースに記憶されている連関度を参照し、新たに物理的計測手段を介して計測された細胞の計測情報に基づいて、上記生物学的計測情報を探索するとともに、探索した上記生物学的計測情報により上記細胞を評価する評価手段とを備えることを特徴とする。 The cell evaluation system according to the present invention is a cell evaluation system that physically measures one or more cells from a group of cells and evaluates the cells, and physically measures the one or more cells. A database in which three or more levels of association between measurement means, measurement information measured by the physical measurement means, and biological measurement information for evaluating cells are stored in advance, and stored in the database. And searching for the biological measurement information based on the measurement information of the cell newly measured through the physical measurement means with reference to the association degree, and the cell based on the searched biological measurement information. And an evaluation means for evaluating.
　また、本発明を適用した細胞評価システムは、細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価システムにおいて、流路中に流下する計測対象の細胞群に対して、構造化された励起光又は照明光を照射し、個々の細胞について時系列的に上記励起光又は上記照明光と相互作用させ、当該細胞の光学的な空間情報を時系列波形にマッピングすることで物理的に計測する物理的計測手段と、上記物理的計測手段により計測された計測情報と、細胞特定情報との３段階以上の連関度が予め記憶されているデータベースと、上記データベースに記憶されている連関度を参照し、新たに物理的計測手段を介して計測された細胞の計測情報に基づいて、細胞特定情報を特定することで評価する評価手段とを備えることを特徴とする。 The cell evaluation system to which the present invention is applied is a cell evaluation system that physically measures one or more cells from a cell group and evaluates the cells. On the other hand, structured excitation light or illumination light is irradiated, and the individual cells interact with the excitation light or illumination light in time series, and the optical spatial information of the cells is mapped to time series waveforms. A physical measurement unit that physically measures, a database in which three or more levels of association between the measurement information measured by the physical measurement unit and the cell identification information are stored in advance, and the database An evaluation means for referring to the stored association degree and evaluating by specifying the cell identification information based on the measurement information of the cell newly measured through the physical measurement means. To.
　本発明に係る細胞評価方法は、細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価方法において、物理的に計測した計測情報と、細胞を評価するための生物学的計測情報にそれぞれが紐付けられた参照用計測情報との３段階以上の連関度を予めデータベースに記憶させ、上記データベースに記憶されている連関度を参照し、新たに計測された細胞の計測情報に基づいて、上記参照用計測情報を探索するとともに、探索した上記参照用計測情報に紐付けられた生物学的計測情報により上記細胞を評価することを特徴とする。 The cell evaluation method according to the present invention is a cell evaluation method in which one or more cells are physically measured from a cell group and the cells are evaluated. Newly measured cells are stored in advance in the database with three or more degrees of association with the reference measurement information associated with the biological measurement information, and the association degree stored in the database is referred to. The reference measurement information is searched based on the measurement information, and the cell is evaluated by biological measurement information linked to the searched reference measurement information.
　本発明に係る細胞評価方法は、細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価方法において、流路中に流下する計測対象の細胞群に対して構造化された励起光又は照明光を照射し、個々の細胞について時系列的に上記励起光と相互作用させ、当該細胞の光学的な空間情報を時系列波形にマッピングすることで物理的に計測し、上記計測した計測情報と、細胞特定情報との３段階以上の連関度を予めデータベースに記憶させ、上記データベースに記憶されている連関度を参照し、新たに計測された細胞の計測情報に基づいて、細胞特定情報を特定することで評価することを特徴とする。 The cell evaluation method according to the present invention is a cell evaluation method in which one or more cells are physically measured from a cell group and the cells are evaluated. Is measured physically by irradiating the excited excitation light or illumination light, interacting with the excitation light in time series for each cell, and mapping the optical spatial information of the cell to the time series waveform. Based on the measurement information of the newly measured cell, the degree of association of three or more levels of the measured measurement information and the cell identification information is stored in the database in advance and the association degree stored in the database is referred to. Thus, evaluation is performed by specifying cell specifying information.
　本発明に係る細胞評価プログラムは、細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価プログラムにおいて、データベースに予め記憶されている、細胞を評価するための生物学的計測情報にそれぞれが紐付けられた参照用計測情報との３段階以上の連関度を参照し、新たに計測された細胞の計測情報に基づいて、上記参照用計測情報を探索するとともに、探索した上記参照用計測情報に紐付けられた生物学的計測情報により上記細胞を評価することをコンピュータに実行させることを特徴とする。 The cell evaluation program according to the present invention is a cell evaluation program for physically measuring one or more cells from a cell group and evaluating the cells. And the reference measurement information associated with the measurement information for each reference linked to the biological measurement information, and searching for the reference measurement information based on the newly measured cell measurement information, The computer is caused to evaluate the cell based on the biological measurement information linked to the searched measurement information for reference.
　本発明に係る細胞評価プログラムは、細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価プログラムにおいて、流路中に流下する計測対象の細胞群に対して構造化された励起光又は照明光を照射し、個々の細胞について時系列的に上記励起光と相互作用させ、当該細胞の光学的な空間情報を時系列波形にマッピングすることで物理的に計測し、上記計測した計測情報と、細胞特定情報との３段階以上の連関度を予めデータベースに記憶させ、上記データベースに記憶されている連関度を参照し、新たに計測された細胞の計測情報に基づいて、細胞特定情報を特定することで評価する評価することをコンピュータに実行させることを特徴とする。 The cell evaluation program according to the present invention is a cell evaluation program that physically measures one or more cells from a cell group and evaluates the cells. Is measured physically by irradiating the excited excitation light or illumination light, interacting with the excitation light in time series for each cell, and mapping the optical spatial information of the cell to the time series waveform. Based on the measurement information of the newly measured cell, the degree of association of three or more levels of the measured measurement information and the cell identification information is stored in the database in advance and the association degree stored in the database is referred to. Thus, the evaluation is performed by specifying the cell specifying information and causing the computer to execute the evaluation.
　本発明を適用した細胞評価システムによれば、イメージングフローサイトメーター分析部を介して新たに取得した細胞の計測情報から、上述した連関度を参照することで参照用計測情報を選択し、更にこの参照用計測情報から生物学的計測情報を介して細胞種を判別し、ひいては細胞を評価することが可能となる。しかも本発明を適用した細胞評価システムによれば、これらの評価動作を人手を介することなく自動的に行うことが可能となる。これにより、新たに取得した細胞の計測情報に基づいて、観察者の主観的判断によらず、生物学的に正しい正解情報に基づく細胞情報の分類、評価を高速かつ正確に行うことが可能となる。またイメージングフローサイトメーター分析部によるイメージングフローサイトメトリー技術により特定される個々の細胞について何れも自動的に細胞種を特定し、ひいては細胞を評価することも可能となることから、当該技術が生み出す膨大な情報量を有効に活用することが可能となる。 According to the cell evaluation system to which the present invention is applied, reference measurement information is selected by referring to the association degree described above from cell measurement information newly acquired via the imaging flow cytometer analysis unit. It becomes possible to discriminate the cell type from the reference measurement information via the biological measurement information and to evaluate the cell. In addition, according to the cell evaluation system to which the present invention is applied, it is possible to automatically perform these evaluation operations without human intervention. This makes it possible to quickly and accurately classify and evaluate cell information based on biologically correct correct information regardless of the observer's subjective judgment based on newly acquired cell measurement information. Become. In addition, it is possible to automatically identify cell types for individual cells identified by the imaging flow cytometry technology by the imaging flow cytometer analysis unit, and thus to evaluate the cells. It is possible to effectively use a large amount of information.
第１実施形態における細胞評価システムのブロック構成図である。It is a block block diagram of the cell evaluation system in 1st Embodiment. イメージングフローサイトメーター分析部の詳細な構成を示す図である。It is a figure which shows the detailed structure of an imaging flow cytometer analysis part. 計測対象の細胞を峻別する方法について説明するための図である。It is a figure for demonstrating the method of discriminating the measurement object cell. データベースに記憶されている、計測情報と、参照用計測情報との３段階以上の連関度を示す図である。It is a figure which shows the 3 or more steps of correlation of measurement information and the reference measurement information which are memorize | stored in the database. 選択した参照用計測情報から細胞種を特定するまでのプロセスを示す図である。It is a figure which shows the process until it specifies a cell type from the selected measurement information for reference. 参照用計測情報を選択することで、生物学的計測情報を介して即座に細胞種を特定する例を示す図である。It is a figure which shows the example which specifies a cell type immediately via biological measurement information by selecting the measurement information for a reference. 計測情報と、生物学的計測情報との３段階以上の連関度を示す図である。It is a figure which shows the relevance of 3 steps | paragraphs or more of measurement information and biological measurement information. 第２実施形態における細胞評価システムのブロック構成図である。It is a block block diagram of the cell evaluation system in 2nd Embodiment. 第２実施形態における学習済みモデルの例を示す図である。It is a figure which shows the example of the learned model in 2nd Embodiment. 細胞特定情報毎に時系列波形をそれぞれタグ付けして学習させる例を示す図である。It is a figure which shows the example trained by tagging a time-sequential waveform for every cell specific information, respectively. 複数の細胞特定情報を混合させた混合体についての時系列波形と各細胞特定情報との３段階以上の連関度を通じて学習させる例を示す図である。It is a figure which shows the example made to learn through the 3 or more steps of linkage of the time-sequential waveform about the mixture which mixed several cell specific information, and each cell specific information. 検知すべき細胞特定情報を特定する上でポジティブな時系列波形と、ネガティブな時系列波形に基づき学習済みモデルを構築する例を示す図である。It is a figure which shows the example which builds a learned model based on a positive time series waveform and negative time series waveform in specifying the cell specific information which should be detected. ポジティブな兆候とネガティブな兆候を細胞特定情報と共に取得することで学習済みモデルを作る例を示す図である。It is a figure which shows the example which makes a learned model by acquiring a positive sign and a negative sign with cell specific information.
　以下、本発明を適用した細胞評価システムについて、図面を参照しながら詳細に説明をする。 Hereinafter, a cell evaluation system to which the present invention is applied will be described in detail with reference to the drawings.
第１実施形態
　図１は、本発明を適用した細胞評価システム１の第１実施形態を示している。細胞評価システム１は、光源２と、光源２からの光が照射されるイメージングフローサイトメーター分析部３と、イメージングフローサイトメーター分析部３からの光情報を受光する受光部４と、光源２、イメージングフローサイトメーター分析部３及び受光部４にそれぞれ接続される制御部５と、光源２、受光部４及び制御部５にそれぞれ接続される評価部６と、評価部６、受光部４に接続されるデータベース７とを備えている。 First Embodiment FIG. 1 shows a first embodiment of a cell evaluation system 1 to which the present invention is applied. The cell evaluation system 1 includes a light source 2, an imaging flow cytometer analysis unit 3 that is irradiated with light from the light source 2, a light receiving unit 4 that receives optical information from the imaging flow cytometer analysis unit 3, A control unit 5 connected to the imaging flow cytometer analysis unit 3 and the light receiving unit 4, an evaluation unit 6 connected to the light source 2, the light receiving unit 4 and the control unit 5, and an evaluation unit 6 and a light receiving unit 4, respectively. The database 7 is provided.
　光源２は、イメージングフローサイトメーター分析部３における分析に必要な光を発光する。イメージングフローサイトメーター分析部３は、イメージングフローサイトメトリー法に基づき、個々の細胞を流体中に分散させ、その流体を微細に流下させて光学的に分析し、細胞の形態情報を取得する。このイメージングフローサイトメーター分析部３の詳細は、後段において詳述する。 The light source 2 emits light necessary for analysis in the imaging flow cytometer analysis unit 3. Based on the imaging flow cytometry method, the imaging flow cytometer analysis unit 3 disperses individual cells in a fluid, finely flows down the fluid, and optically analyzes the cells to obtain cell shape information. Details of the imaging flow cytometer analysis unit 3 will be described later.
　受光部４は、イメージングフローサイトメーター分析部３によるイメージングフローサイトメトリー法を介して得られる光情報を受光するためのセンサで構成される。制御部５は、光源２、イメージングフローサイトメーター分析部３、受光部４、評価部６を制御するための中央制御ユニットとしての役割を担う。この制御部５は、例えばパーソナルコンピュータ（ＰＣ）、携帯端末、スマートフォン、ウェアラブル端末、タブレット型端末等で構成される。 The light receiving unit 4 includes a sensor for receiving optical information obtained through the imaging flow cytometry method performed by the imaging flow cytometer analyzing unit 3. The control unit 5 serves as a central control unit for controlling the light source 2, the imaging flow cytometer analysis unit 3, the light receiving unit 4, and the evaluation unit 6. This control part 5 is comprised with a personal computer (PC), a portable terminal, a smart phone, a wearable terminal, a tablet type terminal etc., for example.
　評価部６は、受光部４から得られた光情報を取得し、更にデータベースに記憶されている情報を参照し、イメージングフローサイトメーター中を流下する細胞を評価する。この細胞の評価は、細胞種の判別、細胞の機能や特性等が含まれる。この評価部６も同様に、ＰＣ、携帯端末、スマートフォン、ウェアラブル端末、タブレット型端末等で構成される。ちなみに、この評価部６は、上述した制御部５と同一デバイスで構成されていてもよい。データベース７は、上述した評価部６による細胞の評価を行う上で必要な情報を記憶するためのハードディスク等で構成されている。 The evaluation unit 6 acquires the optical information obtained from the light receiving unit 4, and further refers to the information stored in the database to evaluate the cells flowing down in the imaging flow cytometer. This cell evaluation includes cell type discrimination, cell function and characteristics, and the like. Similarly, the evaluation unit 6 includes a PC, a mobile terminal, a smartphone, a wearable terminal, a tablet terminal, and the like. Incidentally, this evaluation part 6 may be comprised with the same device as the control part 5 mentioned above. The database 7 is composed of a hard disk or the like for storing information necessary for evaluating cells by the evaluation unit 6 described above.
　図２は、イメージングフローサイトメーター分析部３の詳細な構成を示している。このイメージングフローサイトメーター分析部３は、流路３１中を流れる個々の細胞８につき、３次元的な蛍光画像を取得する。このイメージングフローサイトメーター分析部３は、第１対物レンズ３２と、第１ビームスプリッタ３３と、第１レンズ３４と、第２レンズ３５と、第２ビームスプリッタ３６と、第２対物レンズ３８と、ミラー３９と、第３レンズ４０と、第４レンズ４１とを備えている。 FIG. 2 shows a detailed configuration of the imaging flow cytometer analysis unit 3. The imaging flow cytometer analysis unit 3 acquires a three-dimensional fluorescence image for each cell 8 flowing in the flow path 31. The imaging flow cytometer analysis unit 3 includes a first objective lens 32, a first beam splitter 33, a first lens 34, a second lens 35, a second beam splitter 36, a second objective lens 38, A mirror 39, a third lens 40, and a fourth lens 41 are provided.
　第１対物レンズ３２は、光源２から供給されてくる光を流路３１中に焦点を合わせ、また流路３１から反射する戻り光を集光して、これを第１ビームスプリッタ３３へと送る。第１ビームスプリッタ３３は、光源２から発光された光をそのまま通過させると共に、第１対物レンズ３２からの戻り光を反射させる役割を担う。第１レンズ３４、第２レンズ３５は、第１ビームスプリッタ３３を反射した戻り光の焦点位置やスポット径の調整等を行う。第２ビームスプリッタ３６は、第２レンズ３５を通過した戻り光の一部を第２対物レンズ３８に向けて反射させると共に、戻り光の一部をそのまま通過させる。第２対物レンズ３８は、第２ビームスプリッタ３６を反射してくる戻り光をミラー３９へ導き、またミラー３９から反射してきた戻り光を集光し、これについて第２ビームスプリッタ３６を通過させて第４レンズ４１へと導く。ミラー３９は、第２対物レンズ３８からの戻り光を反射する。第３レンズ４０は、第２ビームスプリッタ３６を通過した戻り光を集光して受光部４を構成する第１受光センサ４ａに結像させる。第４レンズ４１は、第２ビームスプリッタ３６を通過した戻り光を集光して受光部４を構成する第２受光センサ４ｂに結像させる。なお、このイメージングフローサイトメーター分析部３の各構成要素は、制御部５による制御に基づいて動作することとなる。このイメージングフローサイトメーター分析部３は、ＵＳ２０１５／０１９２７６７Ａ１の開示技術を流用するものであってもよい。 The first objective lens 32 focuses the light supplied from the light source 2 into the flow path 31, collects the return light reflected from the flow path 31, and sends this to the first beam splitter 33. . The first beam splitter 33 plays a role of allowing the light emitted from the light source 2 to pass through as it is and reflecting the return light from the first objective lens 32. The first lens 34 and the second lens 35 adjust the focal position and spot diameter of the return light reflected from the first beam splitter 33. The second beam splitter 36 reflects a part of the return light that has passed through the second lens 35 toward the second objective lens 38 and passes a part of the return light as it is. The second objective lens 38 guides the return light reflected from the second beam splitter 36 to the mirror 39, collects the return light reflected from the mirror 39, and passes this through the second beam splitter 36. Guide to the fourth lens 41. The mirror 39 reflects the return light from the second objective lens 38. The third lens 40 condenses the return light that has passed through the second beam splitter 36 and forms an image on the first light receiving sensor 4 a that constitutes the light receiving unit 4. The fourth lens 41 condenses the return light that has passed through the second beam splitter 36 and forms an image on the second light receiving sensor 4 b that constitutes the light receiving unit 4. Each component of the imaging flow cytometer analysis unit 3 operates based on control by the control unit 5. The imaging flow cytometer analysis unit 3 may be one that uses the disclosed technology of US2015 / 0192767A1.
　第１受光センサ４ａは、結像された戻り光の光情報に基づいて、細胞８のｘ－ｙ平面画像を撮像する。第２受光センサ４ｂは、結像された戻り光の光情報に基づいて、細胞８のｘ－ｚ平面画像を撮像する。第１受光センサ４ａ、第２受光センサ４ｂは、それぞれ受光した各平面画像の光情報を電気信号に変換して、これを評価部６やデータベース７へ送信する。第１受光センサ４ａ、第２受光センサ４ｂは、それぞれ制御部５による制御に基づいて動作することとなる。 The first light receiving sensor 4a captures an xy plane image of the cell 8 based on the optical information of the formed return light. The second light receiving sensor 4b captures an xz plane image of the cell 8 based on the optical information of the formed return light. The first light receiving sensor 4a and the second light receiving sensor 4b convert the received light information of each planar image into an electric signal and transmit it to the evaluation unit 6 and the database 7. The first light receiving sensor 4a and the second light receiving sensor 4b operate based on control by the control unit 5, respectively.
　次に、上述した構成からなる細胞評価システム１の動作について説明をする。イメージングフローサイトメーター分析部３は、上述した構成に基づき、流路３１中を流れる一以上の細胞を物理的に計測する。この細胞８の物理的な計測は、可視画像、電磁波、蛍光、位相、透過、分光、多色、散乱、反射、コヒーレントラマン、ラマン又は吸収・散乱・透過・蛍光スペクトル、音、テラヘルツ、インピーダンスの何れかを介して行う。以下では、蛍光画像を３次元的に計測する場合を例にとり説明をする。 Next, the operation of the cell evaluation system 1 having the above-described configuration will be described. The imaging flow cytometer analysis unit 3 physically measures one or more cells flowing in the flow channel 31 based on the above-described configuration. The physical measurement of the cell 8 includes visible image, electromagnetic wave, fluorescence, phase, transmission, spectroscopy, multicolor, scattering, reflection, coherent Raman, Raman or absorption / scattering / transmission / fluorescence spectrum, sound, terahertz, impedance Through either. Hereinafter, a case where a fluorescence image is measured three-dimensionally will be described as an example.
　先ず計測対象の細胞８の種類、形態は、本発明の効果を妨げない限り特に限定されるものではなく、目的に応じて計測対象の細胞を選択することができる。したがって、本発明の計測対象としての細胞８は、浮遊細胞であってもよく、接着細胞であってもよい。明細胞には、細菌（単細胞生物）も含まれる。また２以上の細胞を計測対象してもよく、かかる場合には、複数の細胞があつまった塊（スフェロイド等）を計測対象としてもよい。 First, the type and form of the cell 8 to be measured are not particularly limited as long as the effects of the present invention are not hindered, and the cell to be measured can be selected according to the purpose. Therefore, the cell 8 as the measurement target of the present invention may be a floating cell or an adherent cell. Bright cells also include bacteria (single cell organisms). Two or more cells may be measured, and in such a case, a mass (spheroid or the like) in which a plurality of cells are collected may be measured.
　また計測対象の細胞８を峻別する上では、図３に示すように、少なくとも一つの細胞８と少なくとも一つのビーズ９とを備えたコンパートメント９１を準備し、各コンパートメント９１における細胞８の３次元的な蛍光画像のイメージング情報とビーズ９のイメージング情報とを共に測定することを経時的に繰り返すようにしてもよい。各コンパートメント９１におけるビーズ９のイメージング情報を各コンパートメント９１における細胞８を特定する上での指標とする。 In order to distinguish the cells 8 to be measured, as shown in FIG. 3, compartments 91 including at least one cell 8 and at least one bead 9 are prepared, and the three-dimensional cell 8 in each compartment 91 is prepared. It may be possible to repeat measurement of both the imaging information of the fluorescent image and the imaging information of the beads 9 over time. The imaging information of the beads 9 in each compartment 91 is used as an index for specifying the cell 8 in each compartment 91.
　計測対象の細胞８の数は、シングルセル解析の観点からは、１コンパートメント９１当たり１つであることが望ましいが、これに限定されるものではなく、１コンパートメント９１当たりの細胞８の数を複数としてもよい。ｔ＝０（min）からｔ＝１０（min）まで時系列的に細胞８が変態し、或いは動的挙動を示す場合においても、ビーズ９を指標とすることで、そのコンパートメント９１を特定することができ、ひいてはこれに含まれる細胞８を特定することが可能となる。その結果、一つの細胞８に着目した時系列的な動的変化データを連続的に取得することが可能となる。 The number of cells 8 to be measured is preferably one per compartment 91 from the viewpoint of single cell analysis, but is not limited to this, and the number of cells 8 per compartment 91 is plural. It is good. Even when the cell 8 is transformed in time series from t = 0 (min) to t = 10 (min) or exhibits dynamic behavior, the compartment 91 is specified by using the beads 9 as an index. As a result, it becomes possible to specify the cell 8 contained in the cell. As a result, time-series dynamic change data focusing on one cell 8 can be obtained continuously.
　流路３１を流れる細胞８に対して、光源２からの可視域の光が第１対物レンズ３２から照射され、その戻り光は、第１対物レンズ３２により集光され、第１ビームスプリッタ３３を反射し、第１レンズ３４、３５を通過する。更にこの戻り光の一部は、第２ビームスプリッタ３６をそのまま通過して第３レンズ４０を介して第１受光センサ４ａに結像される。またこの戻り光の他の一部は、第２ビームスプリッタ３６を反射し、ミラー３９を反射した上で、第４レンズ４１を介して第２受光センサ４ｂに結像される。 Visible light from the light source 2 is irradiated from the first objective lens 32 to the cells 8 flowing through the flow path 31, and the return light is collected by the first objective lens 32, and passes through the first beam splitter 33. Reflected and passes through the first lenses 34 and 35. Further, a part of the return light passes through the second beam splitter 36 as it is, and forms an image on the first light receiving sensor 4 a via the third lens 40. Further, another part of the return light is reflected by the second beam splitter 36, reflected by the mirror 39, and imaged on the second light receiving sensor 4b via the fourth lens 41.
　その結果、第１受光センサ４ａにより細胞８のｘ－ｙ平面画像が撮像され、第２受光センサ４ｂにより細胞８のｘ－ｚ平面画像が撮像される結果、この細胞８の３次元的な蛍光画像が形成されることとなる。細胞８が流路３１中の第１対物レンズ３２による撮像範囲を通過する過程で、この撮像動作を連続して行うことにより、細胞８の３次元的な全体像の蛍光画像の光情報を取得することも可能となる。 As a result, an xy plane image of the cell 8 is captured by the first light receiving sensor 4a, and an xz plane image of the cell 8 is captured by the second light receiving sensor 4b. As a result, the three-dimensional fluorescence of the cell 8 is captured. An image will be formed. In the process where the cell 8 passes through the imaging range of the first objective lens 32 in the flow path 31, the imaging information is continuously performed to obtain optical information of the fluorescence image of the three-dimensional whole image of the cell 8. It is also possible to do.
　このようにして第１受光センサ４ａ、第２受光センサ４ｂを介して得られた計測情報は、評価部６、データベース７へと送られる。 The measurement information obtained through the first light receiving sensor 4 a and the second light receiving sensor 4 b in this way is sent to the evaluation unit 6 and the database 7.
　なお上述した動作例は、細胞８の３次元的な蛍光画像を物理的な計測情報として取得する場合について説明をしたが、仮に２次元的な平面画像を物理的な計測情報として取得する場合には、第１受光センサ４ａによるｘ－ｙ平面画像、第２受光センサ４ｂによるｘ－ｚ平面画像の何れかを取得すればよい。 In the above-described operation example, the case where a three-dimensional fluorescence image of the cell 8 is acquired as physical measurement information has been described. However, when a two-dimensional planar image is acquired as physical measurement information, In this case, either the xy plane image by the first light receiving sensor 4a or the xz plane image by the second light receiving sensor 4b may be acquired.
　評価部６は、第１受光センサ４ａ、第２受光センサ４ｂから送られてきた細胞８の物理的な計測情報を解析することにより、細胞８を新たに評価する。この評価部６は、この評価の過程において、データベース７に記憶されている情報を参照する。 The evaluation unit 6 newly evaluates the cell 8 by analyzing physical measurement information of the cell 8 sent from the first light receiving sensor 4a and the second light receiving sensor 4b. The evaluation unit 6 refers to information stored in the database 7 in the course of this evaluation.
　このデータベース７には、図４に示すように、予め計測情報と、参照用計測情報との３段階以上の連関度が予め記憶されている。 As shown in FIG. 4, the database 7 stores in advance three or more levels of association between measurement information and reference measurement information.
　上述したイメージングフローサイトメーター分析部３により撮像された計測情報としての細胞の画像がこの連関度を介して左側に配列し、参照用計測情報がこの連関度を介して右側に配列している。参照用計測情報は、いわゆる細胞の画像の典型例を羅列したものである。この参照用計測情報は、細胞の形態上の特徴に応じて、例えば画像Ａ～Ｋに分類されている。 The cell images as measurement information imaged by the imaging flow cytometer analysis unit 3 described above are arranged on the left side through this association degree, and the reference measurement information is arranged on the right side through this association degree. The reference measurement information is a list of typical examples of so-called cell images. This reference measurement information is classified into, for example, images AK according to the morphological characteristics of the cells.
　この連関度は、新たにイメージングフローサイトメーター分析部３等を通じて撮像された計測情報としての細胞の画像が、この参照用計測情報として予め細胞の形態毎に定義された画像Ａ～Ｋの何れと関連性が高いかを示すものである。また計測情報としての細胞の画像も図４に示すように画像Ｐ１１～Ｐ１６に予め類型化されている。このように計測情報と参照用計測情報は、予め画像毎に類型化されており、その類型化された計測情報の画像Ｐ１１～Ｐ１６と、参照用計測情報の画像Ａ～Ｋとが互いに3段階以上の連関度を介し
て互いに関連付けられている。この連関度は、換言すれば、計測情報の画像が、いかなる参照用計測情報の画像に紐付けられる可能性が高いかを示す指標であり、計測情報から参照用計測情報を選択する上での的確性を示すものである。 The degree of association is determined based on whether a cell image as measurement information newly imaged through the imaging flow cytometer analysis unit 3 or the like is used as any of the images A to K previously defined for each cell form as the reference measurement information. It indicates whether the relevance is high. Further, the cell image as the measurement information is also categorized in advance into images P11 to P16 as shown in FIG. In this way, the measurement information and the reference measurement information are classified in advance for each image, and the classified measurement information images P11 to P16 and the reference measurement information images A to K are in three stages. They are associated with each other through the association described above. In other words, the degree of association is an index indicating whether there is a high possibility that the image of the measurement information is associated with the image of the reference measurement information. In selecting the reference measurement information from the measurement information, It shows the accuracy.
　例えば、計測情報の画像Ｐ１１は、参照用計測情報の画像Ａと連関度８０％、画像Ｂと連関度６０％、画像Ｈと連関度４０％、画像Ｊと連関度２０％であることが示されている。同様に計測情報の画像Ｐ１３は、参照用計測情報の画像Ｄと連関度１００％、画像Ｅと連関度８０％、画像Ｆと連関度６０％であることが示されている。同様に計測情報の画像Ｐ１５は、参照用計測情報の画像Ｈと連関度８０％、画像Ｄと連関度４０％、画像Ａと連関度２０％であることが示されている。ちなみに図４中の計測情報と参照用計測情報との間をつなぐ線の太さがその連関度の大きさを示しており、計測情報と参照用計測情報との間で線が連結されていない場合には、連関度が０％であることを意味している。 For example, the measurement information image P11 has an association degree of 80% with the reference measurement information image A, an association degree with the image B of 60%, an association degree with the image H of 40%, and an association degree with the image J of 20%. Has been. Similarly, it is shown that the measurement information image P13 has an association degree of 100% with the reference measurement information image D, an association degree of the image E of 80%, and an association degree of the image F of 60%. Similarly, it is shown that the measurement information image P15 has an association degree of 80% with the reference measurement information image H, an association degree with the image D of 40%, and an association degree with the image A of 20%. Incidentally, the thickness of the line connecting between the measurement information and the reference measurement information in FIG. 4 indicates the magnitude of the degree of association, and the line is not connected between the measurement information and the reference measurement information. In this case, it means that the degree of association is 0%.
　なお連関度は、いわゆる機械学習を通じて更新が可能なモデルで構成されていてもよく、ニューラルネットワークで構成されていてもよい。またこの連関度は、深層学習がなされることを前提としたネットワークで構成されていてもよい。 Note that the association degree may be constituted by a model that can be updated through so-called machine learning, or may be constituted by a neural network. In addition, this association degree may be configured by a network on the assumption that deep learning is performed.
　評価部６は、このようにデータベース７に記憶されている連関度を参照し、受光部４（第１受光センサ４ａ、第２受光センサ４ｂ）を介して新たに取得した細胞８の計測情報が何れの参照用計測情報に該当する可能性が高いかを判別する。受光部４を介して新たに取得した細胞８の計測情報が、計測情報において予め累計した画像Ｐ１４に類似するのであれば、この画像Ｐ１４と連関度の高い参照用計測情報の画像Ｇに最も該当する可能性が高いことを判別し、次に画像Ｃに該当する可能性が高いことを判別する。受光部４を介して新たに取得した細胞８の計測情報が、計測情報において予め累計した画像Ｐ１６に類似するのであれば、この画像Ｐ１６と連関度の高い参照用計測情報の画像Ｊに最も該当する可能性が高いことを判別し、次に画像Ｋに該当する可能性が高いことを判別する。 The evaluation unit 6 refers to the association degree stored in the database 7 in this manner, and the measurement information of the cell 8 newly acquired via the light receiving unit 4 (first light receiving sensor 4a, second light receiving sensor 4b) is obtained. It is determined which reference measurement information is likely to correspond. If the measurement information of the cell 8 newly acquired via the light receiving unit 4 is similar to the image P14 accumulated in advance in the measurement information, this image P14 and the reference measurement information image G having a high degree of association are most applicable. It is determined that there is a high possibility of the image C, and next, it is determined that there is a high possibility that the image C is applicable. If the measurement information of the cell 8 newly acquired via the light receiving unit 4 is similar to the image P16 accumulated in advance in the measurement information, this image P16 and the reference measurement information image J having a high degree of association are most applicable. It is determined that there is a high possibility of the image K, and next, it is determined that there is a high possibility of corresponding to the image K.
　評価部６は、受光部４を介して新たに取得した細胞８の計測情報に基づき、これら連関度を参照することにより、参照用計測情報の画像を選択する作業を行う。このとき、評価部６は、受光部４を介して新たに取得した細胞８の計測情報と最も連関度の高い参照用計測情報の画像を選択するようにしてもよい。上述したように連関度が高いほど、その選択の的確性が高くなるためである。しかし、評価部６は、最も連関度の高い参照用計測情報の画像を選択する場合に限定されることはなく、連関度が中程度のもの、又は連関度が低いものをあえて選択するようにしてもよい。また、これ以外に計測情報と参照用計測情報との間で矢印が繋がっていない連関度が０％である参照用計測情報を選択してもよいことは勿論である。ちなみに評価部６は、この参照用計測情報を一つ選択する場合に限定されるものではなく、連関度を参照した上であえて複数の参照用計測情報を選択するようにしてもよい。 The evaluation unit 6 performs an operation of selecting an image of the reference measurement information by referring to these association degrees based on the measurement information of the cell 8 newly acquired through the light receiving unit 4. At this time, the evaluation unit 6 may select an image of reference measurement information having the highest degree of association with the measurement information of the cell 8 newly acquired via the light receiving unit 4. This is because, as described above, the higher the degree of association, the higher the accuracy of the selection. However, the evaluation unit 6 is not limited to the case of selecting the reference measurement information image having the highest degree of association. The evaluation unit 6 may intentionally select one having a medium degree of association or a low degree of association. May be. In addition to this, it is needless to say that reference measurement information having an association degree of 0% in which no arrow is connected between the measurement information and the reference measurement information may be selected. Incidentally, the evaluation unit 6 is not limited to the case where one piece of reference measurement information is selected, and a plurality of pieces of reference measurement information may be selected after referring to the association degree.
　次に評価部６は、選択した参照用計測情報から細胞種を特定する。図５は、選択した参照用計測情報から細胞種を特定するまでのプロセスを示している。各参照用計測情報は、それぞれ生物学的計測情報に紐付けられている。この参照用計測情報と生物学的計測情報は、通常１対１で紐付けられているが、これに限定されるものではなく、１対複数、複数対１、又は複数対複数で互いに紐付けられていてもよい。 Next, the evaluation unit 6 identifies the cell type from the selected reference measurement information. FIG. 5 shows a process until the cell type is specified from the selected reference measurement information. Each reference measurement information is associated with biological measurement information. The reference measurement information and the biological measurement information are usually linked one-to-one. However, the reference measurement information and the biological measurement information are not limited to this. One-to-multiple, multiple-to-one, or multiple-to-multiple are linked to each other. It may be done.
　この生物学的計測情報は、細胞種を特定するために、例えば、トランスクリプトーム、ゲノム、エピゲノム、タンパク質、代謝物、糖、脂質、細胞の時系列的変化傾向等を介して計測された情報である。この生物学的計測情報は、参照用計測情報を構成する画像Ａ～Ｋの細胞毎に予めトランスクリプトームやゲノム等を介して取得された情報である。画像Ａ～Ｋの各細胞についてこの生物学的計測情報を取得することにより、その画像Ａ～Ｋの各細胞の正確な細胞の種類を把握することが可能となる。各生物学的計測情報とそこから特定される細胞種は通常１：１で対応する。この図５の例では、参照用計測情報の画像Ａの細胞の生物学的計測情報を取得した結果、「ゲノム○○○」であった場合、その「ゲノム○○○」から細胞種ａを特定することが可能となる。参照用計測情報の画像Ｂの細胞の生物学的計測情報を取得した結果、「脂質の時系列的変化傾向△△△」であった場合、その「脂質の時系列的変化傾向△△△」から細胞種ｂを特定することが可能となる。参照用計測情報の画像Ｋの細胞の生物学的計測情報を取得した結果、「トランスクリプトーム□□□」であった場合、その「トランスクリプトーム□□□」から細胞種ｋを特定することが可能となる。更にこの生物学的計測情報は、より精細または異なる光学的情報を取得する光学イメージング計測情報を含む。つまり、生物学的計測情報は、細胞種等以外に、イメージング（形態）解析結果も含む概念である。 This biological measurement information is information measured via, for example, transcriptome, genome, epigenome, protein, metabolite, sugar, lipid, time-series change tendency of cells, etc. in order to specify the cell type. It is. This biological measurement information is information acquired in advance for each cell of the images AK constituting the reference measurement information via a transcriptome, a genome, or the like. By acquiring this biological measurement information for each cell in the images AK, it is possible to grasp the exact cell type of each cell in the images AK. Each biological measurement information and the cell type identified therefrom usually correspond with 1: 1. In the example of FIG. 5, when the biological measurement information of the cell of the image A for reference measurement information is “genome OO”, the cell type a is determined from the “genome OO”. It becomes possible to specify. As a result of obtaining the biological measurement information of the cell of the image B of the reference measurement information, when it is “a time-series change tendency of lipid ΔΔΔ”, the “trend-time change tendency of lipid ΔΔΔ” From this, it becomes possible to specify the cell type b. If the result of obtaining the biological measurement information of the cell in the reference measurement information image K is “transcriptome □□□”, specify the cell type k from the “transcriptome □□□”. Is possible. Furthermore, this biological measurement information includes optical imaging measurement information for acquiring finer or different optical information. That is, biological measurement information is a concept that includes imaging (morphological) analysis results in addition to cell types and the like.
　データベース７には、この図５に示すような参照用計測情報毎に対応する生物学的計測情報が記憶され、さらにこの生物学的計測情報から特定される細胞種が記憶されている。このため評価部６は、このデータベース７を参照することにより、選択した参照用計測情報に紐付けられる生物学的計測情報を判別し、その判別した生物学的計測情報により特定される細胞種を判別する。 The database 7 stores biological measurement information corresponding to each reference measurement information as shown in FIG. 5, and further stores cell types specified from the biological measurement information. Therefore, the evaluation unit 6 refers to the database 7 to determine the biological measurement information associated with the selected reference measurement information, and selects the cell type specified by the determined biological measurement information. Determine.
　即ち、評価部６は、図６に示すように参照用計測情報を選択することができれば、この生物学的計測情報を介して即座に細胞種を判別することが可能となる。評価部６は、判別した細胞種を図示しないディスプレイ等を始めとした表示部を介して表示する。これによりユーザは、図示しない表示部を視認することにより、新たに取得した細胞８の細胞種を即座に把握することが可能となる。 That is, if the evaluation unit 6 can select the reference measurement information as shown in FIG. 6, the cell type can be immediately discriminated through this biological measurement information. The evaluation unit 6 displays the determined cell type via a display unit such as a display (not shown). Thus, the user can immediately grasp the newly acquired cell type of the cell 8 by visually recognizing a display unit (not shown).
　即ち、本発明を適用した細胞評価システム１によれば、イメージングフローサイトメーター分析部３を介して新たに取得した細胞の計測情報から、上述した連関度を参照することで参照用計測情報を選択し、更にこの参照用計測情報から生物学的計測情報を介して細胞種を判別することが可能となる。しかも本発明を適用した細胞評価システム１によれば、これらの評価動作を人手を介することなく自動的に行うことが可能となる。これにより、物理的計測情報に基づいて、観察者の主観的判断によらず、生物学的に正しい正解情報に基づく細胞情報の分類、評価を高速かつ正確に行うことが可能となる。またイメージングフローサイトメーター分析部３によるイメージングフローサイトメトリー技術により特定される個々の細胞について何れも自動的に細胞種を特定することも可能となることから、当該技術が生み出す膨大な情報量を有効に活用することが可能となる。 That is, according to the cell evaluation system 1 to which the present invention is applied, the reference measurement information is selected by referring to the association degree described above from the cell measurement information newly acquired via the imaging flow cytometer analysis unit 3. In addition, the cell type can be discriminated from the reference measurement information via the biological measurement information. Moreover, according to the cell evaluation system 1 to which the present invention is applied, these evaluation operations can be automatically performed without manual intervention. This makes it possible to quickly and accurately classify and evaluate cell information based on biologically correct correct information, based on physical measurement information, regardless of the observer's subjective judgment. In addition, since it is possible to automatically specify the cell type for each individual cell specified by the imaging flow cytometry technique by the imaging flow cytometer analysis unit 3, the enormous amount of information generated by the technique is effective. It becomes possible to utilize it.
　また、本発明を適用した細胞評価システム１では、３段階以上に設定されている連関度を介して参照用計測情報の探索を行う点に特徴がある。連関度は、例えば０～１００％までの数値で記述することができるが、これに限定されるものではなく３段階以上の数値で記述できるものであればいかなる段階で構成されていてもよい。 Further, the cell evaluation system 1 to which the present invention is applied is characterized in that the reference measurement information is searched through the association degree set in three or more stages. The association degree can be described by a numerical value of, for example, 0 to 100%, but is not limited to this, and may be configured at any stage as long as it can be described by a numerical value of three or more levels.
　このような３段階以上の数値で表される連関度に基づいて探索することで、複数の参照用計測情報が選ばれる状況下において、当該連関度の高い順に探索して表示することも可能となる。このように連関度の高い順にユーザに表示できれば、より可能性の高い参照用計測情報を優先的に選択し、ひいてはそこから特定される細胞種を表示することもできる。一方、連関度の低い参照用計測情報であってもセカンドオピニオンという意味で、そこから特定される細胞種を表示することができ、ファーストオピニオンにおいて表示された細胞種が理解できない場合等において有用性を発揮することができる。 By searching based on the degree of association represented by numerical values of three or more levels, it is possible to search and display in order of the degree of association in a situation where a plurality of reference measurement information is selected. Become. In this way, if it can be displayed to the user in the descending order of relevance, it is possible to preferentially select reference measurement information with a higher possibility, and to display a cell type specified therefrom. On the other hand, even if it is measurement information for reference with low relevance, it is possible to display the cell type identified from it in the sense of a second opinion, which is useful when the cell type displayed in the first opinion cannot be understood. Can be demonstrated.
　これに加えて、本発明によれば、連関度が１％のような極めて低い参照用計測情報も見逃すことなく判断することができる。連関度が極めて低い参照用計測情報であっても僅かな兆候として繋がっているものであり、何十回、何百回に一度は、参照用計測情報として役に立つ場合もあることをユーザに対して注意喚起することができる。 In addition to this, according to the present invention, it is possible to make a determination without missing an extremely low reference measurement information having a relevance of 1%. Even if the measurement information for reference is extremely low, it is connected as a small sign, and it is useful to the user that it may be useful as measurement information for reference once every tens or hundreds of times. You can call attention.
　更に本発明によれば、このような３段階以上の連関度に基づいて探索を行うことにより、閾値の設定の仕方で、探索方針を決めることができるメリットがある。閾値を低くすれば、上述した連関度が１％のものであっても漏れなく拾うことができる反面、正解の可能性が低い参照用計測情報に基づく細胞種を沢山拾ってしまう場合もある。一方、閾値を高くすれば、正解の可能性が高い参照用計測情報により特定される細胞種のみ絞り込むことができる反面、何十回、何百回に一度は好適な解を表示する参照用計測情報により特定される細胞種を見落としてしまう場合もある。いずれに重きを置くかは、ユーザ側、システム側の考え方に基づいて決めることが可能となるが、このような重点を置くポイントを選ぶ自由度を高くすることが可能となる。 Furthermore, according to the present invention, there is an advantage that a search policy can be determined in a manner of setting a threshold by performing a search based on such three or more levels of association. If the threshold value is lowered, it can be picked up without omission even if the association degree is 1%, but there are cases where many cell types are picked up based on the reference measurement information with a low possibility of correct answer. On the other hand, if the threshold value is increased, only the cell types specified by the reference measurement information that is highly likely to be correct can be narrowed down. On the other hand, the reference measurement that displays a suitable solution once every tens or hundreds of times. The cell type specified by the information may be overlooked. It is possible to decide which to place importance on the basis of the idea on the user side and the system side, but it is possible to increase the degree of freedom in selecting points to place such emphasis.
　また、本発明によれば、図３の計測情報について、各計測情報Ｐ１１～Ｐ１６それぞれに対して参照用計測情報の連関度が予め設定されている場合を例にとり説明をしたが、これに限定されるのではない。計測情報の組み合わせと、参照用計測情報との３段階以上の連関度が予め記憶され、新たに物理的に計測された２以上の計測情報に基づいて評価を行うようにしてもよい。 Further, according to the present invention, the measurement information in FIG. 3 has been described by taking as an example the case where the association degree of the reference measurement information is set in advance for each of the measurement information P11 to P16. It is not done. Three or more levels of association between the combination of measurement information and the reference measurement information may be stored in advance, and evaluation may be performed based on two or more measurement information newly physically measured.
　かかる場合には、例えば計測情報Ｐ１１と、Ｐ１２の組み合わせに対して、参照用計測情報Ｄの連関度が６０％、参照用計測情報Ｅの連関度が４０％等のように定義されているものとする。或いは計測情報Ｐ１３、Ｐ１４、Ｐ１５の３つの組み合わせに対して参照用計測情報Ｈの連関度が２０％、参照用計測情報Ｉの連関度が５０％、参照用計測情報Ｊの連関度が７０％等のように定義されている。 In this case, for example, for the combination of the measurement information P11 and P12, the association degree of the reference measurement information D is defined as 60%, the association degree of the reference measurement information E is defined as 40%, and the like. And Alternatively, for the three combinations of the measurement information P13, P14, and P15, the relevance of the reference measurement information H is 20%, the relevance of the reference measurement information I is 50%, and the relevance of the reference measurement information J is 70%. Etc. are defined.
　新たに取得した２以上の計測情報が仮に計測情報Ｐ１１と、計測情報Ｐ１２であった場合、このような連関度を参照すると、参照用計測情報Ｄの連関度が６０％、参照用計測情報Ｅの連関度が４０％等である旨を判別することができる。これら連関度を参照して参照用計測情報を同様に選択することとなる。同様に、新たに取得した２以上の計測情報が仮に計測情報Ｐ１３と、計測情報Ｐ１４と、計測情報Ｐ１５であった場合、このような連関度を参照すると、参照用計測情報Ｈの連関度が２０％、参照用計測情報Ｉの連関度が５０％、参照用計測情報Ｊの連関度が７０％である旨を判別することができる。これら連関度を参照して参照用計測情報を同様に選択することとなる。 If two or more newly acquired pieces of measurement information are the measurement information P11 and the measurement information P12, referring to such an association degree, the association degree of the reference measurement information D is 60%, and the reference measurement information E It can be determined that the degree of association is 40% or the like. The reference measurement information is similarly selected with reference to these degrees of association. Similarly, if two or more newly acquired pieces of measurement information are the measurement information P13, the measurement information P14, and the measurement information P15, referring to such an association degree, the association degree of the reference measurement information H is It can be determined that 20%, the association degree of the reference measurement information I is 50%, and the association degree of the reference measurement information J is 70%. The reference measurement information is similarly selected with reference to these degrees of association.
　この計測情報の組み合わせと、参照用計測情報との３段階以上の連関度を参照する例は、特に細胞の撮像画像の集合間において比較検証を行う、いわゆるマルチインスタンス学習等に適用するようにしてもよい。 The example of referring to the three or more levels of association between the combination of measurement information and the reference measurement information is applied to so-called multi-instance learning or the like that performs comparison verification between a set of captured images of cells. Also good.
　かかる場合には、ある疾患に関する細胞画像をそれぞれ参照用計測情報として記憶しておく。次に、ある患者Ａの複数の細胞それぞれに対して計測情報としての細胞画像を複数撮像し、連関度を参照することで参照用計測情報に相当する疾患に該当するか否かを判別するようにしてもよい。 In such a case, each cell image related to a certain disease is stored as reference measurement information. Next, a plurality of cell images as measurement information are captured for each of a plurality of cells of a patient A, and it is determined whether or not the disease corresponds to the reference measurement information by referring to the association degree. It may be.
　本発明では、上述した連関度を更新させるようにしてもよい。つまり、図４に示すような計測情報と、参照用計測情報を随時更新していく。この更新は、例えばインターネットを始めとした公衆通信網を介して提供された情報を反映させるようにしてもよい。また、専門家による研究データや論文、学会発表や、新聞記事、書籍等の内容に基づいてシステム側又はユーザ側が人為的に、又は自動的に更新するようにしてもよい。これらの更新処理においては人工知能を活用するようにしてもよい。 In the present invention, the association degree described above may be updated. That is, measurement information as shown in FIG. 4 and reference measurement information are updated as needed. This update may reflect information provided via a public communication network such as the Internet. Further, the system side or the user side may be updated artificially or automatically based on the contents of research data, papers, conference presentations, newspaper articles, books, etc. by experts. Artificial intelligence may be used in these update processes.
　この連関度の更新は、計測情報と、参照用計測情報との関係性に関する情報が入る都度、連関度を上昇させ、或いは下降させる。例えば、ある計測情報の画像が、ある参照用計測情報の画像に対応していることが論文や学会発表、その他実験的検証に基づく研究データ等を通じて新たに確認できた場合、その計測情報の画像と参照用計測情報の画像との連関度を上昇させる。また、ある計測情報の画像が、ある参照用計測情報の画像に対応していないことが論文や学会発表、その他実験的検証に基づく研究データ等を通じて新たに確認できた場合、その計測情報の画像と参照用計測情報の画像との連関度を下降させる。 This renewal of the association increases or decreases the association every time information about the relationship between the measurement information and the reference measurement information is entered. For example, if it is newly confirmed that a certain measurement information image corresponds to a certain reference measurement information image through research data based on a paper, a conference presentation, or other experimental verification, the measurement information image And the degree of association between the reference measurement information image and the image of the reference measurement information is increased. In addition, if it is newly confirmed that a measurement information image does not correspond to a reference measurement information image through research papers based on papers, conference presentations, or other experimental verification, the measurement information image And the relevance of the reference measurement information image is lowered.
　連関度は、上述したように３段階以上とされていることで、このような連関度の上昇させたい場合、又は下降させたい場合に自在に対応することが可能となる。この連関度の更新そのものを上述した機械学習、深層学習を通じて行うようにしてもよい。 As described above, the degree of relevance is set to three or more levels, so that it is possible to freely deal with cases where it is desired to increase or decrease the degree of relevance. This relevancy update itself may be performed through the above-described machine learning and deep learning.
　さらに今までに無い計測情報が新たに撮像された場合、これが実はある参照用計測情報に対応していることが実験的に確認できた場合には、これらの間に新たに連関度を設定するようにしてもよい。そして、この新たな計測情報が当該参照用計測情報に対応している報告が上がる都度、この新たに設定した連関度を徐々に上昇させるようにしてもよい。 Furthermore, when measurement information that has never existed has been newly imaged, if it has been experimentally confirmed that this actually corresponds to some reference measurement information, a new association degree is set between them. You may do it. Then, each time the report that the new measurement information corresponds to the reference measurement information increases, the newly set association degree may be gradually increased.
　なお、上述した実施の形態においては、細胞種を特定するための生物学的計測情報にそれぞれが紐付けられた参照用計測情報を選択する場合を例にとり説明した。本発明は、この細胞種を特定する例に限定されるものではなく、細胞の機能や特性を評価する場合に適用するようにしてもよい。かかる場合には、図５に示す生物学的計測情報により特定されるものが細胞種ではなく、細胞の機能や細胞の特性となる。かかる場合においても同様にそれぞれの生物学的計測情報に対する、細胞の各機能や細胞の各特性の対応関係を予め取得し、データベース７に記憶させておくこととなる。その結果、図６に示すように、参照用計測情報の画像が選択されると、この生物学的計測情報を介して即座に細胞種に代替される細胞の各機能や細胞の各特性が即座に決定されることとなる。 In the embodiment described above, the case where reference measurement information associated with biological measurement information for specifying a cell type is selected as an example has been described. The present invention is not limited to the example of specifying the cell type, and may be applied when evaluating the function and characteristics of the cell. In such a case, what is specified by the biological measurement information shown in FIG. 5 is not a cell type but a cell function or a cell characteristic. In such a case as well, the correspondence relationship of each cell function and each cell characteristic with respect to each biological measurement information is acquired in advance and stored in the database 7. As a result, as shown in FIG. 6, when an image of reference measurement information is selected, each function or characteristic of the cell that is immediately replaced by the cell type is immediately determined through this biological measurement information. Will be determined.
　また、上述した実施の形態においては、計測情報を２次元又は３次元の蛍光画像を介して取得する場合を例にとり説明したが、他の電磁波、明視野、暗視野、蛍光、位相、透過、分光、多色、散乱、反射、コヒーレントラマン、ラマン又は吸収・散乱・透過・蛍光スペクトル、音、テラヘルツ、インピーダンスの何れかにより細胞８を物理的に計測し、そこから計測情報を得る場合も同様である。かかる場合には、参照用計測情報もその物理的計測手段に合わせて予め典型例を用意しておき、連関度で互いの関連付けを行っておく。 In the above-described embodiment, the case where measurement information is acquired via a two-dimensional or three-dimensional fluorescence image has been described as an example, but other electromagnetic waves, bright field, dark field, fluorescence, phase, transmission, The same applies when cell 8 is physically measured by spectroscopy, multicolor, scattering, reflection, coherent Raman, Raman or absorption / scattering / transmission / fluorescence spectrum, sound, terahertz, impedance, and measurement information is obtained from it. It is. In such a case, the reference measurement information is also prepared in advance according to the physical measurement means, and is associated with each other with the degree of association.
　例えば物理的計測手段が蛍光スペクトル分析である場合には、計測情報、参照用計測情報共に蛍光スペクトルとなる。参照用計測情報は、各波長に対するピーク位置や変化傾向等が予め類型化されている。この類型化された参照用計測情報としての蛍光スペクトルパターンに対して、細胞８から新たに取得した蛍光スペクトルからなる計測情報に基づいて、連関度を参照することで参照用計測情報を選択することとなる。 For example, when the physical measurement means is fluorescence spectrum analysis, both the measurement information and the reference measurement information are fluorescence spectra. In the reference measurement information, the peak position and the change tendency with respect to each wavelength are classified in advance. The reference measurement information is selected by referring to the association degree based on the measurement information including the fluorescence spectrum newly acquired from the cell 8 for the categorized fluorescence spectrum pattern as the reference measurement information. It becomes.
　また物理的な計測手段として、イメージングフローサイトメトリー法以外に、顕微鏡を用いたマイクロウェル内のコンパートメント観察法を含めた顕微鏡を用いる観察法、イメージングサイトメトリー法等、いかなる方法に基づくものであってもよい。また液体中を流下する細胞を物理的に計測する場合以外に、液体中には無い細胞を計測する場合においても本発明を適用するようにしてもよい。使用する顕微鏡は、光学顕微鏡（明視野、位相差、蛍光、共焦点レーザー、ラマン等を含む）、または電子顕微鏡（透過型、走査型）等であってもよい。 In addition to the imaging flow cytometry method, the physical measurement means can be based on any method such as an observation method using a microscope including a compartment observation method in a microwell using a microscope, an imaging cytometry method, etc. Also good. Further, the present invention may be applied to the case of measuring cells that are not in the liquid in addition to the case of physically measuring the cells flowing down in the liquid. The microscope used may be an optical microscope (including bright field, phase difference, fluorescence, confocal laser, Raman, etc.) or an electron microscope (transmission type, scanning type).
　また、本発明では、図７に示すように、予め計測情報と、生物学的計測情報との３段階以上の連関度が予め記憶させるようにしてもよい。 Further, in the present invention, as shown in FIG. 7, the degree of association of three or more levels of measurement information and biological measurement information may be stored in advance.
　上述したイメージングフローサイトメーター分析部３により撮像された計測情報としての細胞の画像がこの連関度を介して左側に配列し、生物学的計測情報がこの連関度を介して右側に配列している。生物学的計測情報は、例えばそれぞれ細胞種（細胞の機能や特性等）を特定するための各種ゲノム情報等が羅列されている。この生物学的計測情報は、ゲノム情報等に応じて予め類型化されており、図７に示すようにゲノム情報Ｏ～ゲノム情報Ｙ等が分類されている。 The image of the cell as measurement information imaged by the imaging flow cytometer analysis unit 3 described above is arranged on the left side via this association degree, and the biological measurement information is arranged on the right side via this association degree. . The biological measurement information includes, for example, various types of genome information for specifying cell types (cell functions, characteristics, etc.). This biological measurement information is categorized in advance according to genome information and the like, and genome information O to genome information Y and the like are classified as shown in FIG.
　この連関度は、新たにイメージングフローサイトメーター分析部３等を通じて撮像された計測情報としての細胞の画像が、この参照用計測情報として予め分類したゲノム情報Ｏ～ゲノム情報Ｙ等の何れと関連性が高いかを示すものであり、互いに３段階以上の連関度を介して互いに関連付けられている。この連関度は、換言すれば、計測情報の画像が、いかなる生物学的計測情報に紐付けられる可能性が高いかを示す指標であり、計測情報から生物学的計測情報を選択する上での的確性を示すものである。 This degree of association is related to any of the genome information O to genome information Y, etc., that the cell image as measurement information newly imaged through the imaging flow cytometer analysis unit 3 etc. is previously classified as this reference measurement information Are related to each other through three or more levels of association. In other words, the degree of association is an index indicating which biological measurement information is likely to be associated with an image of measurement information. In selecting biological measurement information from measurement information, It shows the accuracy.
　例えば、計測情報の画像Ｐ１１は、生物学的計測情報のゲノム情報Ｏと連関度８０％、ゲノム情報Ｐと連関度６０％、ゲノム情報Ｖと連関度４０％、ゲノム情報Ｘと連関度２０％であることが示されている。同様に計測情報の画像Ｐ１３は、生物学的計測情報のゲノム情報Ｒと連関度１００％、ゲノム情報Ｓと連関度８０％、ゲノム情報Ｔと連関度６０％であることが示されている。ちなみに図７中の計測情報と生物学的計測情報との間をつなぐ線の太さがその連関度の大きさを示しており、計測情報と参照用計測情報との間で線が連結されていない場合には、連関度が０％であることを意味している。 For example, the measurement information image P11 has an association degree of 80% with the genome information O of the biological measurement information, an association degree of 60% with the genome information P, an association degree of 40% with the genome information V, and an association degree of 20% with the genome information X. It is shown that. Similarly, it is shown that the measurement information image P13 has an association degree of 100% with the genome information R of the biological measurement information, an association degree of 80% with the genome information S, and an association degree of 60% with the genome information T. Incidentally, the thickness of the line connecting the measurement information and the biological measurement information in FIG. 7 indicates the magnitude of the association, and the line is connected between the measurement information and the reference measurement information. If not, it means that the relevance is 0%.
　なお、この連関度も同様に、いわゆる機械学習を通じて更新が可能なモデルで構成されていてもよく、ニューラルネットワークで構成されていてもよい。またこの連関度は、深層学習がなされることを前提としたネットワークで構成されていてもよい。 Note that this association degree may also be constituted by a model that can be updated through so-called machine learning, or may be constituted by a neural network. In addition, this association degree may be configured by a network on the assumption that deep learning is performed.
　評価部６は、このようにデータベース７に記憶されている連関度を参照し、受光部４（第１受光センサ４ａ、第２受光センサ４ｂ）を介して新たに取得した細胞８の計測情報が何れの生物学的計測情報に該当する可能性が高いかを判別する。受光部４を介して新たに取得した細胞８の計測情報が、計測情報において予め累計した画像Ｐ１４に類似するのであれば、この画像Ｐ１４と連関度の高い生物学的計測情報のゲノム情報Ｕに最も該当する可能性が高いことを判別し、次にゲノム情報Ｑに該当する可能性が高いことを判別する。受光部４を介して新たに取得した細胞８の計測情報が、計測情報において予め累計した画像Ｐ１６に類似するのであれば、この画像Ｐ１６と連関度の高い生物学的計測情報のゲノム情報Ｘに最も該当する可能性が高いことを判別し、次にゲノム情報Ｙに該当する可能性が高いことを判別する。 The evaluation unit 6 refers to the association degree stored in the database 7 in this manner, and the measurement information of the cell 8 newly acquired via the light receiving unit 4 (first light receiving sensor 4a, second light receiving sensor 4b) is obtained. It is determined which biological measurement information is likely to correspond. If the measurement information of the cell 8 newly acquired via the light receiving unit 4 is similar to the image P14 accumulated in advance in the measurement information, the genome information U of the biological measurement information highly related to the image P14 is included. It is determined that there is a high possibility that it corresponds to the most, and next, it is determined that there is a high possibility that it corresponds to the genome information Q. If the measurement information of the cell 8 newly acquired via the light receiving unit 4 is similar to the image P16 accumulated in advance in the measurement information, the genome information X of the biological measurement information highly related to the image P16 is used. It is determined that there is a high possibility that it corresponds to the most, and then it is determined that there is a high possibility that it corresponds to the genome information Y.
　評価部６は、受光部４を介して新たに取得した細胞８の計測情報に基づき、これら連関度を参照することにより、生物学的計測情報を選択する作業を行う。このとき、評価部６は、受光部４を介して新たに取得した細胞８の計測情報と最も連関度の高い生物学的計測情報を選択するようにしてもよい。上述したように連関度が高いほど、その選択の的確性が高くなるためである。しかし、評価部６は、最も連関度の高い生物学的計測情報の画像を選択する場合に限定されることはなく、連関度が中程度のもの、又は連関度が低いものをあえて選択するようにしてもよい。また、これ以外に計測情報と生物学的計測情報との間で矢印が繋がっていない連関度が０％である生物学的計測情報を選択してもよいことは勿論である。ちなみに評価部６は、この生物学的計測情報を一つ選択する場合に限定されるものではなく、連関度を参照した上であえて複数の参照用計測情報を選択するようにしてもよい。 The evaluation unit 6 performs an operation of selecting biological measurement information by referring to these association degrees based on the measurement information of the cell 8 newly acquired via the light receiving unit 4. At this time, the evaluation unit 6 may select biological measurement information having the highest degree of association with the measurement information of the cell 8 newly acquired through the light receiving unit 4. This is because, as described above, the higher the degree of association, the higher the accuracy of the selection. However, the evaluation unit 6 is not limited to selecting an image of biological measurement information with the highest degree of association, and it may be selected with a medium degree of association or a low degree of association. It may be. In addition to this, it is needless to say that biological measurement information having an association degree of 0% in which no arrow is connected between measurement information and biological measurement information may be selected. Incidentally, the evaluation unit 6 is not limited to selecting one biological measurement information, but may select a plurality of reference measurement information after referring to the association degree.
　次に評価部６は、選択した生物学的計測情報から細胞種を特定する。生物学的計測情報はそもそも細胞種と関連付けられているものであることから、この生物学的計測情報を選択することができれば、即座に細胞種を特定することができる。また細胞種以外の細胞の機能や特性の評価と生物学的計測情報とを予め関連付けておくことにより、細胞種以外の細胞の評価を行うことも可能となる。 Next, the evaluation unit 6 identifies the cell type from the selected biological measurement information. Since the biological measurement information is originally associated with the cell type, if this biological measurement information can be selected, the cell type can be identified immediately. It is also possible to evaluate cells other than the cell type by associating in advance the evaluation of the function and characteristics of the cells other than the cell type with the biological measurement information.
　このような図７に示す計測情報と、生物学的計測情報との連関度についても更新するようにしてもよい。この更新は、例えばインターネットを始めとした公衆通信網を介して提供された情報を反映させるようにしてもよい。また、専門家による研究データや論文、学会発表や、新聞記事、書籍等の内容に基づいてシステム側又はユーザ側が人為的に、又は自動的に更新するようにしてもよい。これらの更新処理においては人工知能を活用するようにしてもよい。 Such association between the measurement information shown in FIG. 7 and the biological measurement information may be updated. This update may reflect information provided via a public communication network such as the Internet. Further, the system side or the user side may be updated artificially or automatically based on the contents of research data, papers, conference presentations, newspaper articles, books, etc. by experts. Artificial intelligence may be used in these update processes.
　この連関度の更新は、計測情報と、生物学的計測情報との関係性に関する情報が入る都度、連関度を上昇させ、或いは下降させる。例えば、ある計測情報の画像が、ある生物学的計測情報のゲノム情報に対応していることが論文や学会発表、その他実験的検証に基づく研究データ等を通じて新たに確認できた場合、その計測情報の画像と生物学的計測情報のゲノム情報との連関度を上昇させる。また、ある計測情報の画像が、ある生物学的計測情報のゲノム情報に対応していないことが論文や学会発表、その他実験的検証に基づく研究データ等を通じて新たに確認できた場合、その計測情報の画像と生物学的計測情報のゲノム情報との連関度を下降させる。 This renewal of association increases or decreases the association every time information about the relationship between measurement information and biological measurement information is entered. For example, if an image of certain measurement information can be newly confirmed through research papers based on papers, conference presentations, or other experimental verifications, the measurement information corresponds to the genomic information of a certain biological measurement information. Increase the degree of association between the image and the genome information of biological measurement information. In addition, if it is newly confirmed that a measurement information image does not correspond to the genomic information of a certain biological measurement information through a research paper based on a paper, a conference presentation, or other experimental verification, the measurement information The degree of association between the image and the genome information of biological measurement information is lowered.
　第２実施形態
　以下、第２実施形態に係る細胞評価システム１´について図面を参照しながら詳細に説明をする。この第２実施形態において、上述した第１実施形態と同一の構成要素、部材に関しては、同一の符号を付すことにより以下での説明を省略する。 Second Embodiment Hereinafter, a cell evaluation system 1 ′ according to a second embodiment will be described in detail with reference to the drawings. In the second embodiment, the same components and members as those in the first embodiment described above are denoted by the same reference numerals, and the description thereof will be omitted.
　図８は、本発明を適用した細胞評価システム１´の形態を示している。細胞評価システム１´は、イメージングフローサイトメーター分析部３´と、イメージングフローサイトメーター分析部３に接続されたレンズ７２と、レンズ７２から光が導かれるＰＭＴ（一画素素子）７１とを備えている。またこの細胞評価システム１´は、ＰＭＴ７１により検知されたデータを評価する評価部６と、評価部６に対してそれぞれ接続された制御部７及びデータベース７とを備えている。 FIG. 8 shows a form of a cell evaluation system 1 ′ to which the present invention is applied. The cell evaluation system 1 ′ includes an imaging flow cytometer analysis unit 3 ′, a lens 72 connected to the imaging flow cytometer analysis unit 3, and a PMT (one pixel element) 71 from which light is guided from the lens 72. Yes. The cell evaluation system 1 ′ includes an evaluation unit 6 that evaluates data detected by the PMT 71, and a control unit 7 and a database 7 that are connected to the evaluation unit 6, respectively.
　この既知静止ランダム構造照明部３０は、特定の濃淡パターンを介して励起光を発光する、いわゆる構造化照明を行う光構造である。この既知静止ランダム構造照明部３０は、圧縮センシングによる設計方法、又はこれ以外のいかなる設計方法に基づいて設計される。このような光構造を通じて構造化された励起光又は照明光を照射することができ、サンプル面に光パターンを形成させることができる。このような既知静止ランダム構造照明部３０中の流路中に計測対象となる細胞８を含む細胞群が流下する。その結果、流路中に流下する計測対象の細胞群に対して励起光が照射されることになる。個々の細胞８は、回折素子などの光学素子を通り、時系列的に励起光と相互作用することになる。当該細胞８の光学的な空間情報（形態に関する情報）を時系列波形にマッピングすることで物理的な計測を実現することができる。イメージングフローサイトメーター分析部３´は、いわゆるゴーストサイトメトリー法を適用するようにしてもよい。また得られた波形信号に基いて細胞の画像を再構成するようにしてもよいが、この細胞画像の再構成は、この第２実施形態において特段必須ではない。 The known stationary random structure illumination unit 30 is an optical structure that performs so-called structured illumination that emits excitation light through a specific gray pattern. This known stationary random structure illumination unit 30 is designed based on a design method using compressed sensing or any other design method. The structured excitation light or illumination light can be irradiated through such an optical structure, and an optical pattern can be formed on the sample surface. A cell group including the cells 8 to be measured flows down in the flow path in the known stationary random structure illumination unit 30. As a result, the excitation light is irradiated to the measurement target cell group flowing down in the flow path. Each cell 8 passes through an optical element such as a diffraction element and interacts with excitation light in a time series. Physical measurement can be realized by mapping the optical spatial information (information on the form) of the cell 8 to a time-series waveform. The imaging flow cytometer analysis unit 3 ′ may apply a so-called ghost cytometry method. Further, the cell image may be reconstructed based on the obtained waveform signal, but the reconstruction of the cell image is not particularly essential in the second embodiment.
　即ち、励起光は時系列に細胞８の異なる部分を励起し相互作用を起こさせ、細胞８の空間情報は時間情報に圧縮変換されることになる。イメージングフローサイトメーター分析部３´中を流れる細胞８の動きと、既知静止ランダム構造照明部３０にからの励起光（照明光）の発光が進行し、励起光（照明光）と細胞の重なっている部分が変化していき、重なりの総量の時間変化を計測する。つまり、図８に示すような横軸が時間軸であり、縦軸が信号強度軸からなる時系列波形を検出することができる。 That is, the excitation light excites different parts of the cell 8 in time series to cause interaction, and the spatial information of the cell 8 is compressed and converted into time information. The movement of the cell 8 flowing in the imaging flow cytometer analysis unit 3 ′ and the emission of excitation light (illumination light) from the known stationary random structure illumination unit 30 proceed, and the excitation light (illumination light) overlaps with the cells. The part that changes is changed, and the time change of the total amount of overlap is measured. That is, a time series waveform in which the horizontal axis is a time axis and the vertical axis is a signal intensity axis as shown in FIG. 8 can be detected.
　このような時系列波形を高速高感度な一画素素子であるＰＭＴ７１により検出する。レンズ７２は、イメージングフローサイトメーター分析部３´からの光をＰＭＴ７１に対して導き、集光させることにより、一画素分の信号計測を行う。 Such a time series waveform is detected by the PMT 71, which is a single pixel element with high speed and high sensitivity. The lens 72 performs signal measurement for one pixel by guiding the light from the imaging flow cytometer analysis unit 3 ′ to the PMT 71 and collecting it.
　評価部６は、このようにして得られた時系列波形に基づいて、イメージングフローサイトメーター分析部３´を流れる細胞における細胞特定情報を探索する。ここでいう細胞特定情報とは、イメージングフローサイトメーター分析部３´を流れる細胞を特定するためのいかなる情報を含む概念であり、例えば細胞フェノタイプ、細胞特性、細胞種、細胞内分子局在、細胞内構造、細胞形態、細胞成熟度、それら重要な細胞キャラクタリゼーション等を規定するための情報である。以下の例では、この細胞特定情報として、イメージングフローサイトメーター分析部３´を流れる細胞の細胞種を探索する場合を例にとり説明をする。 The evaluation unit 6 searches for cell specific information in the cells flowing through the imaging flow cytometer analysis unit 3 ′ based on the time-series waveform obtained in this way. The cell specifying information here is a concept including any information for specifying a cell flowing through the imaging flow cytometer analysis unit 3 ′. For example, cell phenotype, cell characteristics, cell type, intracellular molecular localization, Information for defining intracellular structures, cell morphology, cell maturity, important cell characterization, and the like. In the following example, a case where the cell type of the cell flowing through the imaging flow cytometer analysis unit 3 ′ is searched as an example of the cell specifying information will be described.
　評価部６は、この細胞特定情報を探索するに当たり、事前に学習済みモデルをデータベース７内に構築して記憶させておく。この学習済みモデルの例を図９に示す。 The evaluator 6 constructs and stores a learned model in the database 7 in advance when searching for this cell identification information. An example of this learned model is shown in FIG.
　図９に示すように、時系列波形がこの連関度を介して左側に配列し、細胞特定情報（細胞種）がこの連関度を介して右側に配列している。連関度については、上述した図４の例と同様であるが、この例では、時系列波形が、細胞特定情報の何れと関連性が高いかを示すものである。 As shown in FIG. 9, the time-series waveform is arranged on the left side through this association degree, and the cell identification information (cell type) is arranged on the right side through this association degree. The degree of association is the same as in the example of FIG. 4 described above, but in this example, the time-series waveform indicates which of the cell identification information is more relevant.
　例えば、時系列波形のＵ１１は、細胞特定情報の細胞種Ａと連関度８０％、細胞種Ｂと連関度６０％、細胞種Ｈと連関度４０％、細胞種Ｊと連関度２０％であることが示されている。同様に時系列波形のＵ１３は、細胞特定情報の細胞種Ｄと連関度１００％、細胞種Ｅと連関度８０％、細胞種Ｆと連関度６０％であることが示されている。同様に時系列波形のＵ１５は、細胞特定情報の細胞種Ｈと連関度８０％、細胞種Ｄと連関度４０％、細胞種Ａと連関度２０％であることが示されている。ちなみに、時系列波形と細胞特定情報との間で線が連結されていない場合には、連関度が０％であることを意味している。 For example, U11 of the time-series waveform has an association degree of 80% with cell type A, an association degree of 60% with cell type B, an association degree of 40% with cell type H, and an association degree of 20% with cell type J. It has been shown. Similarly, U13 of the time-series waveform is shown to have an association degree of 100% with the cell type D of the cell identification information, an association degree of 80% with the cell type E, and an association degree of 60% with the cell type F. Similarly, U15 of the time-series waveform is shown to have an association degree of 80% with the cell type H of the cell identification information, an association degree of 40% with the cell type D, and an association degree of 20% with the cell type A. Incidentally, when the line is not connected between the time-series waveform and the cell specifying information, it means that the association degree is 0%.
　なお連関度は、いわゆる機械学習を通じて更新が可能なモデルで構成されていてもよく、ニューラルネットワークで構成されていてもよい。またこの連関度は、深層学習がなされることを前提としたネットワークで構成されていてもよい。 Note that the association degree may be constituted by a model that can be updated through so-called machine learning, or may be constituted by a neural network. In addition, this association degree may be configured by a network on the assumption that deep learning is performed.
　評価部６は、このようにデータベース７に記憶されている連関度を参照し、ＰＭＴ７１を介して新たに取得した細胞８の時系列波形が何れの細胞特定情報に該当する可能性が高いかを判別する。新たに取得した細胞８の時系列波形が、時系列波形Ｕ１４に類似するのであれば、このＵ１４と連関度の高い細胞特定情報の細胞種Ｇに最も該当する可能性が高いことを判別し、次に細胞種Ｃに該当する可能性が高いことを判別する。新たに取得した細胞８の時系列波形が、Ｕ１６に類似するのであれば、このＵ１６と連関度の高い細胞特定情報の細胞種Ｊに最も該当する可能性が高いことを判別し、次に細胞種Ｋに該当する可能性が高いことを判別する。 The evaluation unit 6 refers to the association degree stored in the database 7 in this manner, and determines which cell specifying information is likely to correspond to the time-series waveform of the cell 8 newly acquired via the PMT 71. Determine. If the newly acquired time-series waveform of the cell 8 is similar to the time-series waveform U14, it is determined that there is a high possibility that the cell type G of the cell identification information having a high degree of association with this U14 is the most relevant. Next, it is determined that there is a high possibility of corresponding to cell type C. If the time series waveform of the newly acquired cell 8 is similar to U16, it is determined that there is a high possibility that the cell type J of the cell identification information having a high degree of association with U16 is most likely, and then the cell It is determined that there is a high possibility that it corresponds to the seed K.
　ちなみに、新たに取得した時系列波形が、データベース７中に記憶させてある学習済みモデルの時系列波形Ｕのいずれに類似するかを判別する際にも、例えばディープラーニングや機械学習といった人工知能の各手法を用いるようにしてもよい。 Incidentally, when it is determined whether the newly acquired time-series waveform is similar to the time-series waveform U of the learned model stored in the database 7, for example, artificial intelligence such as deep learning or machine learning is used. Each method may be used.
　評価部６は、新たに取得した細胞８の時系列情報に基づき、これら連関度を参照することにより、細胞特定情報の何れか１以上を選択する作業を行う。このとき、評価部６は、受光部４を介して新たに取得した細胞８の時系列情報と最も連関度の高い細胞特定情報を選択するようにしてもよい。上述したように連関度が高いほど、その選択の的確性が高くなるためである。しかし、評価部６は、最も連関度の高い細胞特定情報の画像を選択する場合に限定されることはなく、連関度が中程度のもの、又は連関度が低いものをあえて選択するようにしてもよい。また、これ以外に時系列情報と細胞特定情報との間で矢印が繋がっていない連関度が０％である細胞特定情報を選択してもよいことは勿論である。ちなみに評価部６は、この細胞特定情報を一つ選択する場合に限定されるものではなく、連関度を参照した上であえて複数の細胞特定情報を選択するようにしてもよい。 The evaluation unit 6 performs an operation of selecting any one or more of the cell identification information by referring to these association degrees based on the newly acquired time series information of the cells 8. At this time, the evaluation unit 6 may select cell specifying information having the highest degree of association with the time series information of the cell 8 newly acquired via the light receiving unit 4. This is because, as described above, the higher the degree of association, the higher the accuracy of the selection. However, the evaluation unit 6 is not limited to selecting the image of the cell identification information with the highest degree of association. The evaluation unit 6 may select the medium with the medium degree of association or the one with the low degree of association. Also good. In addition to this, it is a matter of course that cell specifying information having an association degree of 0% in which no arrow is connected between the time series information and the cell specifying information may be selected. Incidentally, the evaluation unit 6 is not limited to the case of selecting one piece of the cell specifying information, but may select a plurality of pieces of cell specifying information with reference to the association degree.
　ちなみに、この連関度は、ニューラルネットワークのニューロンに代替されるものであってもよい。かかる場合には、複数の時系列波形の組み合わせに対する１以上の細胞特定情報が連関度を介して関連付けられた学習済みモデルを構築しておく。そして実際の判別時において、上述した方法に基づいて１以上の細胞特定情報を選択するようにしてもよい。 Incidentally, this degree of association may be replaced with a neuron of a neural network. In such a case, a learned model is constructed in which one or more cell identification information for a combination of a plurality of time-series waveforms is associated through association. In actual discrimination, one or more pieces of cell specifying information may be selected based on the method described above.
　図１０は、細胞特定情報毎に時系列波形をそれぞれタグ付けして学習させる例を示している。つまり、各細胞種Ａ、Ｂ毎にそれぞれ時系列波形を学習させる。細胞特定情報毎に時系列波形を３段階以上の連関度を通じて学習させた学習済みモデルを記憶させておく。このような学習済みモデルの例が上述した図９に示すネットワークになる。次にＰＭＴ７１を介して新たに取得した細胞８の時系列波形が何れの細胞特定情報（細胞種Ａ，細胞種Ｂ）に該当するかを上述のように判別する。 FIG. 10 shows an example of learning by tagging time-series waveforms for each cell specifying information. That is, a time-series waveform is learned for each cell type A and B. A learned model in which a time series waveform is learned through three or more levels of association for each cell specifying information is stored. An example of such a learned model is the network shown in FIG. Next, it is determined as described above which cell identification information (cell type A, cell type B) the time series waveform of the cell 8 newly acquired via the PMT 71 corresponds to.
　図１１は、複数の細胞特定情報を混合させた混合体についての時系列波形と各細胞特定情報との３段階以上の連関度を通じて学習させる例を示している。つまり学習済みモデルの構築段階において、細胞種Ａ、細胞種Ｂを混合させた混合体についての時系列波形を取得しておく。その結果得られる時系列波形は、細胞種Ａに応じた兆候と、細胞種Ｂに応じた兆候が共に含まれるものとなっている。このような時系列波形を学習済みモデルとして学習させておき、ＰＭＴ７１を介して計測対象の細胞８が組み合わさった細胞群を新たに取得する。そこから得られる時系列波形から細胞種Ａに応じた兆候、細胞種Ｂに応じた兆候をそれぞれ学習済みモデルと照らし合わせ、何れに該当するか判別する。 FIG. 11 shows an example in which learning is performed through three or more levels of association between a time-series waveform and each cell specifying information about a mixture obtained by mixing a plurality of cell specifying information. That is, in the construction stage of the learned model, a time-series waveform for a mixture obtained by mixing the cell type A and the cell type B is acquired. The time-series waveform obtained as a result includes both a sign corresponding to the cell type A and a sign corresponding to the cell type B. Such a time series waveform is learned as a learned model, and a cell group in which cells 8 to be measured are combined is newly acquired via the PMT 71. From the time-series waveform obtained therefrom, a sign corresponding to the cell type A and a sign corresponding to the cell type B are respectively compared with the learned model to determine which corresponds.
　かかる場合における学習済みモデルは、時系列波形を図９に示す連関度の左側に配列し、この時系列波形は、細胞種Ａに応じた兆候、細胞種Ｂに応じた兆候の何れか一方又は両方の兆候が組み合わさったものとなっている。このような時系列波形に対して、それぞれ細胞特定情報が連関度を介してタグ付けされて学習済みモデルを構成する。ＰＭＴ７１を介して計測対象の細胞８が組み合わさった細胞群から時系列波形を新たに取得し、学習済みモデルを参照した結果、これが図１１に示すように細胞種Ｂの兆候を含むことを判別した場合には、その取得した細胞群において細胞種Ｂが含まれていることを判別することが可能となる。 In the learned model in this case, the time series waveform is arranged on the left side of the association degree shown in FIG. 9, and this time series waveform is either one of the signs according to the cell type A and the signs according to the cell type B or Both signs are a combination. With respect to such time series waveforms, cell specific information is tagged via the association degree to constitute a learned model. As a result of obtaining a new time-series waveform from the cell group in which the cells 8 to be measured are combined via the PMT 71 and referring to the learned model, it is determined that this includes a sign of the cell type B as shown in FIG. In this case, it is possible to determine that the cell type B is included in the acquired cell group.
　図１２は、検知すべき細胞特定情報を特定する上でポジティブな時系列波形と、ネガティブな時系列波形とを予め３段階以上の連関度を通じて学習させた学習済みモデルを構築する例である。図１１の例と同様に、複数の細胞特定情報を混合させた混合体についての時系列波形から、特定の細胞特定情報（例えば細胞種Ｂ）を検出する上で、ポジティブな時系列波形とネガティブな時系列的波形を予め学習させておく。ここでいうポジティブな時系列波形とは、特定の細胞特定情報（例えば細胞種Ｂ）に当てはまる蓋然性が高い波形上の兆候である。一方、ネガティブな時系列波形とは、特定の細胞特定情報（例えば細胞種Ｂ）に当てはまる蓋然性が低い波形上の兆候である。かかる場合には、時系列波形が図９に示す連関度の左側に配列し、ポジティブかネガティブかが連関度の右側に配列し、それぞれ連関度を介して関連付けられている。 FIG. 12 is an example of constructing a learned model in which a positive time series waveform and a negative time series waveform are learned in advance through three or more degrees of association in specifying cell identification information to be detected. Similar to the example of FIG. 11, in detecting specific cell specific information (for example, cell type B) from a time series waveform of a mixture obtained by mixing a plurality of cell specific information, a positive time series waveform and a negative A simple time-series waveform is learned in advance. The positive time-series waveform here is a sign on the waveform having a high probability of being applied to specific cell specific information (for example, cell type B). On the other hand, a negative time-series waveform is a sign on the waveform that has a low probability of being applied to specific cell specific information (for example, cell type B). In such a case, the time series waveform is arranged on the left side of the association degree shown in FIG. 9, and positive or negative is arranged on the right side of the association degree, and each is associated through the association degree.
　そしてＰＭＴ７１を介して計測対象の細胞８が組み合わさった細胞群から時系列波形を新たに取得し、学習済みモデルを参照した結果、ポジティブな時系列波形の兆候が現れた場合には、特定の細胞特定情報（例えば細胞種Ｂ）に当てはまる蓋然性を高くし、ネガティブな時系列波形の兆候が現れた場合には、特定の細胞特定情報（例えば細胞種Ｂ）に当てはまる蓋然性を低くする。つまり、ポジティブな時系列波形に近づくほど検知すべき細胞特定情報を判断する上でプラスに判断し、ネガティブな時系列波形に近づくほど検知すべき細胞特定情報を判断する上でプラスに判断する。最終的にこの特定の細胞特定情報に当てはまるポジティブな兆候とネガティブな兆候を総合的に見極め、特定の細胞特定情報に当てはまる蓋然性を判断していくことになる。 Then, when a time series waveform is newly acquired from the cell group in which the measurement target cells 8 are combined via the PMT 71 and the learned model is referred to, a positive time series waveform sign appears as a result. The probability of being applied to cell specific information (for example, cell type B) is increased, and when a sign of a negative time series waveform appears, the probability of being applied to specific cell specific information (for example, cell type B) is decreased. In other words, the closer to the positive time series waveform, the more positive it is to determine the cell specific information to be detected, and the closer to the negative time series waveform, the more positive it is to determine the cell specific information to be detected. Ultimately, positive signs and negative signs that apply to this specific cell-specific information are comprehensively determined, and the probability of applying to the specific cell-specific information is determined.
　かかる場合において、検知すべき細胞特定情報を特定する上でネガティブな時系列波形のみを予め３段階以上の連関度を通じて学習させた学習済みモデルを記憶させておくようにしてもよい。そしてＰＭＴ７１を介して計測対象の細胞８が組み合わさった細胞群から時系列波形を新たに取得した場合には、ネガティブな時系列波形に該当するか否かを学習済みモデルを参照して判断する。その結果、ネガティブな時系列波形に該当しないものを検知すべき細胞特定情報として特定するようにしてもよい。 In such a case, in order to specify the cell specifying information to be detected, a learned model in which only negative time-series waveforms are learned in advance through three or more levels of association may be stored. When a time series waveform is newly acquired from the cell group in which the cells 8 to be measured are combined via the PMT 71, it is determined with reference to the learned model whether the time series waveform corresponds to the negative time series waveform. . As a result, information that does not correspond to a negative time-series waveform may be specified as cell specifying information to be detected.
　図１２中の「＋」はポジティブな兆候であり、「－」はネガティブな兆候であるが、それぞれの兆候を検出する毎に制御部５による制御を切り替える例を示している。イメージングフローサイトメーター分析部３´の先端が分岐しており、制御部５による制御を通じて細胞を分類できるように構成している場合には、ポジティブな兆候が現れた場合のみその特定の細胞８を分岐させた特定の流路に導くようにしてもよい。かかる場合において制御部５は、電場や磁場等を通じたあらゆる物理的手段を通じて特定の細胞８を導くようにしてもよいことは勿論である。 In FIG. 12, “+” is a positive sign and “−” is a negative sign, but an example in which the control by the control unit 5 is switched every time each sign is detected is shown. In the case where the tip of the imaging flow cytometer analysis unit 3 ′ is branched and configured to be able to classify cells through control by the control unit 5, the specific cell 8 is selected only when a positive sign appears. You may make it guide | induce to the branched specific flow path. In such a case, the control unit 5 may naturally guide the specific cell 8 through any physical means such as an electric field or a magnetic field.
　また図１３は、ポジティブな兆候とネガティブな兆候を細胞特定情報と共に取得することで学習済みモデルを作る例を示している。ポジティブな兆候を呈する時系列波形と共に、緑蛍光生細胞染色と、赤蛍光バイオマーカーを介して得られた細胞特定情報を同時に計測する。同様にネガティブな兆候を呈する時系列波形と共に、緑蛍光生細胞染色と、赤蛍光バイオマーカーを介して得られた細胞特定情報を同時に計測する。 FIG. 13 shows an example of creating a learned model by acquiring positive signs and negative signs together with cell identification information. The cell specific information obtained through the green fluorescent live cell staining and the red fluorescent biomarker is simultaneously measured together with the time-series waveform exhibiting positive signs. Similarly, the cell specific information obtained through the green fluorescent live cell staining and the red fluorescent biomarker is simultaneously measured together with the time-series waveform exhibiting a negative sign.
　このようにて得られた各時系列波形とそれぞれにタグ付けされたポジティブ又はネガティブな細胞特定情報を同時に計測し、学習済みモデルとして蓄積していく。この学習済みモデルを構築する過程において、細胞の蛍光染色または非染色に関わらず、時系列波形と、細胞特定情報の同時計測を行うことが必要となる。 The time-series waveforms obtained in this way and the positive or negative cell specific information tagged to each are measured simultaneously and accumulated as a learned model. In the process of constructing this learned model, it is necessary to simultaneously measure the time-series waveform and the cell specific information regardless of whether the cells are stained or not stained.
　なお、本発明においては、上述のようにして作製した学習済みモデルに基づく教師有り学習データに加え、ポジティブな時系列波形及びネガティブな時系列波形と関連付けられていないラベル無しデータに基づいた半教師有り学習に基づいて、細胞特定情報を特定するようにしてもよいことは勿論である。例えば、血液中のがん細胞のように、大半がどのようにラベルを付けてよいか分からない場合において、一部のみラベルの付け方が既知である場合において有用である。 In the present invention, in addition to supervised learning data based on a learned model created as described above, a semi-teacher based on unlabeled data that is not associated with a positive time-series waveform and a negative time-series waveform Of course, the cell specifying information may be specified based on the learning. For example, it is useful when most of the labeling is known, such as cancer cells in blood, when most do not know how to label.
１　細胞評価システム
２　光源
３　イメージングフローサイトメーター分析部
４　受光部
４ａ　受光センサ
４ｂ　受光センサ
５　制御部
６　評価部
７　データベース
８　細胞
９　ビーズ
３１　流路
３２　第１対物レンズ
３３　第１ビームスプリッタ
３４　第１レンズ
３５　第２レンズ
３６　第２ビームスプリッタ
３８　第２対物レンズ
３９　ミラー
４０　第３レンズ
４１　第４レンズ
７１　ＰＭＴ
７２　レンズ９１　コンパートメント
 
  DESCRIPTION OF SYMBOLS 1 Cell evaluation system 2 Light source 3 Imaging flow cytometer analysis part 4 Light receiving part 4a Light receiving sensor 4b Light receiving sensor 5 Control part 6 Evaluation part 7 Database 8 Cell 9 Bead 31 Flow path 32 1st objective lens 33 1st beam splitter 34 1st Lens 35 Second lens 36 Second beam splitter 38 Second objective lens 39 Mirror 40 Third lens 41 Fourth lens 71 PMT
72 Lens 91 compartment

Claims (20)

  1. 　細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価システムにおいて、
    　上記一以上の細胞を物理的に計測する物理的計測手段と、
    　上記物理的計測手段により計測された計測情報と、細胞を評価するための生物学的計測情報にそれぞれが紐付けられた参照用計測情報との３段階以上の連関度が予め記憶されているデータベースと、
    　上記データベースに記憶されている連関度を参照し、新たに物理的計測手段を介して計測された細胞の計測情報に基づいて、上記参照用計測情報を探索するとともに、探索した上記参照用計測情報に紐付けられた生物学的計測情報により上記細胞を評価する評価手段とを備えること
    　を特徴とする細胞評価システム。     In a cell evaluation system that physically measures one or more cells from a cell group and evaluates the cells,
    Physical measuring means for physically measuring the one or more cells;
    A database that stores in advance three or more levels of association between measurement information measured by the physical measurement means and reference measurement information associated with biological measurement information for evaluating cells. When,
    The reference measurement information is searched while referring to the association degree stored in the database and based on the measurement information of the cell newly measured through the physical measurement means, and the searched measurement information for reference An evaluation means for evaluating the cell based on biological measurement information linked to the cell evaluation system.
  2. 　上記物理的計測手段は、上記一以上の細胞の可視画像、電磁波、蛍光、位相、透過、分光、多色、散乱、反射、コヒーレントラマン、ラマン又は吸収・散乱・透過・蛍光スペクトル、音、テラヘルツ、インピーダンスの何れかを介して物理的に計測し、
    　上記データベースは、上記物理的計測手段による物理的な計測方法に対応した上記参照用計測情報が予め記憶されていること
    　を特徴とする請求項１記載の細胞評価システム。     The physical measurement means includes a visible image of the one or more cells, electromagnetic waves, fluorescence, phase, transmission, spectroscopy, multicolor, scattering, reflection, coherent Raman, Raman or absorption / scattering / transmission / fluorescence spectrum, sound, terahertz , Physically measure through any of the impedance,
    The cell evaluation system according to claim 1, wherein the reference measurement information corresponding to a physical measurement method by the physical measurement unit is stored in the database in advance.
  3. 　上記データベースは、上記計測情報と、トランスクリプトーム、ゲノム、エピゲノム、タンパク質、代謝物、糖、脂質、細胞の時系列的変化傾向の何れかの上記生物学的計測情報にそれぞれが紐付けられた上記参照用計測情報との３段階以上の連関度が予め記憶されていること
    　を特徴とする請求項１又は２記載の細胞評価システム。     The database is associated with the measurement information and the biological measurement information of any one of transcriptome, genome, epigenome, protein, metabolite, sugar, lipid, and time-series change tendency of cells. The cell evaluation system according to claim 1 or 2, wherein a degree of association of three or more stages with the reference measurement information is stored in advance.
  4. 　上記細胞の物理的な計測が、フローサイトメトリー法、顕微鏡を用いたマイクロウェル内のコンパートメント観察法を含めた顕微鏡を用いる観察法、イメージングフローサイトメトリー法の何れかに基づいて行われること
    　を特徴とする請求項１～３のうち何れか１項記載の細胞評価システム。     The physical measurement of the cells is performed based on any one of flow cytometry, observation using a microscope, including a method of observing compartments in a microwell using a microscope, and imaging flow cytometry. The cell evaluation system according to any one of claims 1 to 3.
  5. 　上記データベースは、上記計測情報と、細胞種を特定するための生物学的計測情報にそれぞれが紐付けられた参照用計測情報との３段階以上の連関度が予め記憶され、
    　上記評価手段は、探索した上記参照用計測情報に紐付けられた生物学的計測情報により特定される細胞種を判別すること
    　を特徴とする請求項１～３のうち何れか１項記載の細胞評価システム。     The database stores in advance three or more degrees of association between the measurement information and the reference measurement information associated with the biological measurement information for specifying the cell type,
    The cell according to any one of claims 1 to 3, wherein the evaluation means discriminates a cell type specified by biological measurement information linked to the searched measurement information for reference. Evaluation system.
  6. 　上記データベースは、上記計測情報と、上記参照用計測情報との関係を新たに取得した場合には、これを上記連関度に反映させることで更新すること
    　を特徴とする請求項１～５のうち何れか１項記載の細胞評価システム。     6. The database according to claim 1, wherein when the relationship between the measurement information and the reference measurement information is newly acquired, the database is updated by reflecting the relationship in the association degree. The cell evaluation system according to any one of claims.
  7. 　上記データベースは、上記連関度を機械学習させることで更新すること
    　を特徴とする請求項６記載の細胞評価システム。     The cell evaluation system according to claim 6, wherein the database is updated by machine learning of the association degree.
  8. 　上記データベースは、上記計測情報と、上記参照用計測情報との間で深層学習させることにより３段階以上の連関度が予め記憶されていること
    　を特徴とする請求項６記載の細胞評価システム。     The cell evaluation system according to claim 6, wherein the database stores in advance three or more levels of association by deep learning between the measurement information and the reference measurement information.
  9. 　上記データベースは、上記計測情報の組み合わせと、上記参照用計測情報との３段階以上の連関度が予め記憶され、
    　上記評価手段は、新たに物理的に計測された２以上の計測情報に基づいて評価を行うこと
    　を特徴とする請求項１～６のうち何れか１項記載の細胞評価システム。     The database stores in advance three or more levels of association between the combination of the measurement information and the reference measurement information,
    The cell evaluation system according to any one of claims 1 to 6, wherein the evaluation means performs evaluation based on two or more pieces of measurement information newly physically measured.
  10. 　細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価システムにおいて、
    　上記一以上の細胞を物理的に計測する物理的計測手段と、
    　上記物理的計測手段により計測された計測情報と、細胞を評価するための生物学的計測情報との３段階以上の連関度が予め記憶されているデータベースと、
    　上記データベースに記憶されている連関度を参照し、新たに物理的計測手段を介して計測された細胞の計測情報に基づいて、上記生物学的計測情報を探索するとともに、探索した上記生物学的計測情報により上記細胞を評価する評価手段とを備えること
    　を特徴とする細胞評価システム。     In a cell evaluation system that physically measures one or more cells from a cell group and evaluates the cells,
    Physical measuring means for physically measuring the one or more cells;
    A database in which three or more levels of association between measurement information measured by the physical measurement means and biological measurement information for evaluating cells are stored in advance;
    With reference to the association degree stored in the database, the biological measurement information is searched based on the measurement information of the cells newly measured through the physical measurement means, and the searched biological A cell evaluation system comprising: an evaluation means for evaluating the cell based on measurement information.
  11. 　細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価システムにおいて、
    　流路中に流下する計測対象の細胞群に対して、構造化された励起光又は照明光を照射し、個々の細胞について時系列的に上記励起光又は上記照明光と相互作用させ、当該細胞の光学的な空間情報を時系列波形にマッピングすることで物理的に計測する物理的計測手段と、
    　上記物理的計測手段により計測された計測情報と、細胞特定情報との３段階以上の連関度が予め記憶されているデータベースと、
    　上記データベースに記憶されている連関度を参照し、新たに物理的計測手段を介して計測された細胞の計測情報に基づいて、細胞特定情報を特定することで評価する評価手段とを備えること
    　を特徴とする細胞評価システム。     In a cell evaluation system that physically measures one or more cells from a cell group and evaluates the cells,
    Irradiate structured excitation light or illumination light to the measurement target cell group flowing down in the flow path, and interact with the excitation light or the illumination light in time series for each cell. Physical measurement means for physically measuring optical spatial information by mapping them into time-series waveforms;
    A database in which the degree of association of three or more levels of measurement information measured by the physical measurement means and cell identification information is stored in advance;
    Evaluation means for referring to the degree of association stored in the database and evaluating by specifying the cell identification information based on the measurement information of the cell newly measured through the physical measurement means. Characteristic cell evaluation system.
  12. 　上記データベースには、細胞特定情報毎に時系列波形を３段階以上の連関度を通じて学習させた学習済みモデルを記憶させておくこと
    　を特徴とする請求項１１記載の細胞評価システム。     The cell evaluation system according to claim 11, wherein the database stores a learned model in which a time-series waveform is learned through three or more levels of association for each cell specifying information.
  13. 　上記データベースには、複数の細胞特定情報を混合させた混合体についての時系列波形と各細胞特定情報との３段階以上の連関度を通じて学習させた学習済みモデルを記憶させ、
    　上記評価手段は、新たに物理的計測手段を介して計測された上記混合体の計測情報に基づいて、細胞特定情報を特定することで評価すること
    　を特徴とする請求項１１記載の細胞評価システム。     In the database, a learned model trained through three or more levels of association between a time-series waveform of each mixture obtained by mixing a plurality of cell specific information and each cell specific information is stored,
    The cell evaluation system according to claim 11, wherein the evaluation unit performs evaluation by specifying cell identification information based on measurement information of the mixture newly measured through a physical measurement unit. .
  14. 　上記データベースには、検知すべき細胞特定情報を特定する上でポジティブな時系列波形と、ネガティブな時系列波形とを予め３段階以上の連関度を通じて学習させた学習済みモデルを記憶させ、
    　上記評価手段は、新たに物理的計測手段を介して計測された上記混合体の時系列波形に基づいて、ポジティブな時系列波形に近づくほど検知すべき細胞特定情報を判断する上でプラスに判断し、ネガティブな時系列波形に近づくほど検知すべき細胞特定情報を判断する上でプラスに判断すること
    　を特徴とする請求項１３記載の細胞評価システム。     The database stores a learned model in which positive time-series waveforms and negative time-series waveforms are learned in advance through three or more degrees of association in identifying cell identification information to be detected,
    Based on the time-series waveform of the mixture newly measured through the physical measurement unit, the evaluation unit makes a positive determination in determining cell-specific information that should be detected as it approaches a positive time-series waveform. The cell evaluation system according to claim 13, wherein the cell evaluation system makes a positive determination in determining the cell specifying information to be detected as it approaches a negative time-series waveform.
  15. 　上記データベースには、検知すべき細胞特定情報を特定する上でネガティブな時系列波形を予め３段階以上の連関度を通じて学習させた学習済みモデルを記憶させ、
    　上記評価手段は、新たに物理的計測手段を介して計測された上記混合体の時系列波形に基づいて、ネガティブな時系列波形に該当しないものを検知すべき細胞特定情報として特定すること
    　を特徴とする請求項１３記載の細胞評価システム。     In the database, in order to specify the cell identification information to be detected, a learned model in which a negative time series waveform is learned in advance through three or more levels of association is stored.
    The evaluation means specifies, based on the time series waveform of the mixture newly measured through the physical measurement means, the cell specifying information to be detected that does not correspond to the negative time series waveform. The cell evaluation system according to claim 13.
  16. 　上記評価手段は、上記学習済みモデルに基づく教師有り学習データに加え、上記ポジティブな時系列波形及びネガティブな時系列波形と関連付けられていないラベル無しデータに基づいた半教師有り学習に基づいて、細胞特定情報を特定すること
    　を特徴とする請求項１４記載の細胞評価システム。     In addition to the supervised learning data based on the learned model, the evaluation means is based on semi-supervised learning based on unlabeled data that is not associated with the positive time series waveform and the negative time series waveform. The cell evaluation system according to claim 14, wherein the specific information is specified.
  17. 　細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価方法において、
    　物理的に計測した計測情報と、細胞を評価するための生物学的計測情報にそれぞれが紐付けられた参照用計測情報との３段階以上の連関度を予めデータベースに記憶させ、
    　上記データベースに記憶されている連関度を参照し、新たに計測された細胞の計測情報に基づいて、上記参照用計測情報を探索するとともに、探索した上記参照用計測情報に紐付けられた生物学的計測情報により上記細胞を評価すること
    　を特徴とする細胞評価方法。     In a cell evaluation method for physically measuring one or more cells from a cell group and evaluating the cells,
    Store in the database the degree of association of three or more levels of the measurement information measured physically and the reference measurement information linked to the biological measurement information for evaluating the cells,
    With reference to the association degree stored in the database, the reference measurement information is searched based on the newly measured cell measurement information, and the biology linked to the searched reference measurement information Cell evaluation method, characterized by evaluating the above-mentioned cells based on statistical measurement information.
  18. 　細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価プログラムにおいて、
    　データベースに予め記憶されている、細胞を評価するための生物学的計測情報にそれぞれが紐付けられた参照用計測情報との３段階以上の連関度を参照し、新たに計測された細胞の計測情報に基づいて、上記参照用計測情報を探索するとともに、探索した上記参照用計測情報に紐付けられた生物学的計測情報により上記細胞を評価することをコンピュータに実行させること
    　を特徴とする細胞評価プログラム。     In a cell evaluation program for physically measuring one or more cells from a cell group and evaluating the cells,
    Measurement of newly measured cells with reference to three or more levels of association with reference measurement information linked to biological measurement information for evaluating cells stored in advance in the database A cell characterized in that, based on the information, the reference measurement information is searched, and the computer is caused to evaluate the cell based on the biological measurement information linked to the searched reference measurement information. Evaluation program.
  19. 　細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価方法において、
    　流路中に流下する計測対象の細胞群に対して構造化された励起光又は照明光を照射し、個々の細胞について時系列的に上記励起光と相互作用させ、当該細胞の光学的な空間情報を時系列波形にマッピングすることで物理的に計測し、
    　上記計測した計測情報と、細胞特定情報との３段階以上の連関度を予めデータベースに記憶させ、
    　上記データベースに記憶されている連関度を参照し、新たに計測された細胞の計測情報に基づいて、細胞特定情報を特定することで評価すること
    　を特徴とする細胞評価方法。     In a cell evaluation method for physically measuring one or more cells from a cell group and evaluating the cells,
    Irradiate structured excitation light or illumination light to the measurement target cell group flowing down in the flow path, and cause each cell to interact with the excitation light in a time-series manner. Physically measure by mapping information to time series waveform,
    Store in the database the degree of association of three or more levels of the measured measurement information and the cell identification information,
    A cell evaluation method, characterized by referring to the association degree stored in the database and performing evaluation by specifying cell specific information based on newly measured cell measurement information.
  20. 　細胞群の中から一以上の細胞を物理的に計測してその細胞を評価する細胞評価プログラムにおいて、
    　流路中に流下する計測対象の細胞群に対して構造化された励起光又は照明光を照射し、個々の細胞について時系列的に上記励起光と相互作用させ、当該細胞の光学的な空間情報を時系列波形にマッピングすることで物理的に計測し、
    　上記計測した計測情報と、細胞特定情報との３段階以上の連関度を予めデータベースに記憶させ、
    　上記データベースに記憶されている連関度を参照し、新たに計測された細胞の計測情報に基づいて、細胞特定情報を特定することで評価する評価することをコンピュータに実行させること
    　を特徴とする細胞評価プログラム。
          In a cell evaluation program for physically measuring one or more cells from a cell group and evaluating the cells,
    Irradiate structured excitation light or illumination light to the measurement target cell group flowing down in the flow path, and cause each cell to interact with the excitation light in a time-series manner. Physically measure by mapping information to time series waveform,
    Store in the database the degree of association of three or more levels of the measured measurement information and the cell identification information,
    A cell characterized by causing a computer to perform evaluation by referring to the degree of association stored in the database and identifying cell specific information based on newly measured cell measurement information. Evaluation program.
PCT/JP2018/017499 2017-05-02 2018-05-02 Cell evaluation system and method, and cell evaluation program WO2018203568A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US16/610,481 US11598712B2 (en) 2017-05-02 2018-05-02 System and method for cell evaluation, and cell evaluation program
JP2019515746A JP7176697B2 (en) 2017-05-02 2018-05-02 Cell evaluation system and method, cell evaluation program

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2017091957 2017-05-02
JP2017-091957 2017-05-02

Publications (1)

Publication Number Publication Date
WO2018203568A1 true WO2018203568A1 (en) 2018-11-08

Family

ID=64016144

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2018/017499 WO2018203568A1 (en) 2017-05-02 2018-05-02 Cell evaluation system and method, and cell evaluation program

Country Status (3)

Country Link
US (1) US11598712B2 (en)
JP (1) JP7176697B2 (en)
WO (1) WO2018203568A1 (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2020153946A (en) * 2019-03-22 2020-09-24 シスメックス株式会社 Method for analyzing cells, method for training deep learning algorithm, cell analyzer, device for training deep learning algorithm, cell analysis program, and deep learning algorithm training program
JPWO2021085648A1 (en) * 2019-10-31 2021-05-06
WO2022059468A1 (en) * 2020-09-18 2022-03-24 シスメックス株式会社 Cell analysis method and cell analysis device
WO2022239596A1 (en) * 2021-05-11 2022-11-17 シンクサイト株式会社 Feature amount calculation device, feature amount calculation method, and program
WO2023053574A1 (en) 2021-09-29 2023-04-06 日東紡績株式会社 Method for enriching cells or cell nuclei
EP4246124A1 (en) 2022-03-17 2023-09-20 Sysmex Corporation Specimen analyzer, specimen analysis method, and program

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
MX2021002279A (en) 2018-08-27 2021-05-27 Regeneron Pharma Use of raman spectroscopy in downstream purification.
EP3825663A1 (en) * 2019-11-22 2021-05-26 A·P·E Angewandte Physik & Elektronik GmbH Device for producing light pulses for characterizing, standardizing or calibrating photodetectors in flow cytometers

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016080442A1 (en) * 2014-11-21 2016-05-26 住友電気工業株式会社 Quality evaluation method and quality evaluation device
JP2016189702A (en) * 2015-03-30 2016-11-10 国立大学法人名古屋大学 Cell analysis model creating device and cell analysis model creating method, cell analysis device and cell analysis method, and program
WO2017073737A1 (en) * 2015-10-28 2017-05-04 国立大学法人東京大学 Analysis device

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6249341B1 (en) 1999-01-25 2001-06-19 Amnis Corporation Imaging and analyzing parameters of small moving objects such as cells
JP5534214B2 (en) 2009-10-05 2014-06-25 ベイバイオサイエンス株式会社 Flow cytometer and flow cytometry method
US9823457B2 (en) * 2014-01-08 2017-11-21 The Regents Of The University Of California Multiplane optical microscope
US10900950B2 (en) * 2015-03-30 2021-01-26 National University Corporation Nagoya University Apparatus and analytical evaluation methods using morphological feature parameters of a cell or cell population

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016080442A1 (en) * 2014-11-21 2016-05-26 住友電気工業株式会社 Quality evaluation method and quality evaluation device
JP2016189702A (en) * 2015-03-30 2016-11-10 国立大学法人名古屋大学 Cell analysis model creating device and cell analysis model creating method, cell analysis device and cell analysis method, and program
WO2017073737A1 (en) * 2015-10-28 2017-05-04 国立大学法人東京大学 Analysis device

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP7352365B2 (en) 2019-03-22 2023-09-28 シスメックス株式会社 Cell analysis method, deep learning algorithm training method, cell analysis device, deep learning algorithm training device, cell analysis program, and deep learning algorithm training program
WO2020196074A1 (en) 2019-03-22 2020-10-01 シスメックス株式会社 Cell analysis method, training method for deep learning algorithm, cell analysis device, training method for deep learning algorithm, cell analysis program, and training program for deep learning algorithm
CN113574380A (en) * 2019-03-22 2021-10-29 希森美康株式会社 Cell analysis method, deep learning algorithm training method, cell analysis device, deep learning algorithm training device, cell analysis program, and deep learning algorithm training program
JP2020153946A (en) * 2019-03-22 2020-09-24 シスメックス株式会社 Method for analyzing cells, method for training deep learning algorithm, cell analyzer, device for training deep learning algorithm, cell analysis program, and deep learning algorithm training program
EP3943933A4 (en) * 2019-03-22 2022-12-28 Sysmex Corporation Cell analysis method, training method for deep learning algorithm, cell analysis device, training method for deep learning algorithm, cell analysis program, and training program for deep learning algorithm
JPWO2021085648A1 (en) * 2019-10-31 2021-05-06
WO2021085648A1 (en) * 2019-10-31 2021-05-06 株式会社ニコン Cell evaluation device, program, and cell evaluation method
JP7416084B2 (en) 2019-10-31 2024-01-17 株式会社ニコン Cell evaluation device, program, and cell evaluation method
WO2022059468A1 (en) * 2020-09-18 2022-03-24 シスメックス株式会社 Cell analysis method and cell analysis device
WO2022239596A1 (en) * 2021-05-11 2022-11-17 シンクサイト株式会社 Feature amount calculation device, feature amount calculation method, and program
KR20230170982A (en) 2021-09-29 2023-12-19 니토 보세키 가부시기가이샤 Methods for Enriching Cells or Cell Nuclei
WO2023053574A1 (en) 2021-09-29 2023-04-06 日東紡績株式会社 Method for enriching cells or cell nuclei
EP4246123A1 (en) 2022-03-17 2023-09-20 Sysmex Corporation Specimen analyzer, specimen analysis method, and program
EP4246124A1 (en) 2022-03-17 2023-09-20 Sysmex Corporation Specimen analyzer, specimen analysis method, and program

Also Published As

Publication number Publication date
US11598712B2 (en) 2023-03-07
JP7176697B2 (en) 2022-11-22
JPWO2018203568A1 (en) 2020-04-30
US20200150022A1 (en) 2020-05-14

Similar Documents

Publication Publication Date Title
WO2018203568A1 (en) Cell evaluation system and method, and cell evaluation program
EP3375859B1 (en) Method for constructing classifier, and method for determining life or death of cells using same
US20100135566A1 (en) Analysis and classification, in particular of biological or biochemical objects, on the basis of time-lapse images, applicable in cytometric time-lapse cell analysis in image-based cytometry
JP2024045407A (en) Method and apparatus for detecting entities in body samples
JP2015510592A (en) Flow cytometer with digital holographic microscope
JP6967232B2 (en) Image processing device, image processing method and image processing program
CN110446803A (en) Automatically the cell specified number is collected
WO2017150194A1 (en) Image processing device, image processing method, and program
JP4749637B2 (en) Image analysis method, apparatus, and recording medium
Dunker Hidden secrets behind dots: Improved phytoplankton taxonomic resolution using high‐throughput imaging flow cytometry
JP2022500647A (en) Cell sorting device and method
JP2013109119A (en) Microscope controller and program
Wolf et al. Current approaches to fate mapping and lineage tracing using image data
JP5762764B2 (en) Cell image analysis system
Senft et al. A biologist’s guide to planning and performing quantitative bioimaging experiments
Niederlein et al. Image analysis in high content screening
JP5657910B2 (en) Automatic image analysis using a bright field microscope
US20240118527A1 (en) Fluorescence microscopy for a plurality of samples
Matula et al. Acquiarium: free software for the acquisition and analysis of 3D images of cells in fluorescence microscopy
Richmond et al. Deadnet: Identifying phototoxicity from label-free microscopy images of cells using deep convnets
WO2023042646A1 (en) Classification model generation method, particle determination method, computer program, and information processing device
Hardo et al. Quantitative Microbiology with Microscopy: Effects of Projection and Diffraction
WO2023042647A1 (en) Classification model generation method, particle classification method, computer program, and information processing device
Grimes Image processing and analysis methods in quantitative endothelial cell biology
WO2022269960A1 (en) Particle analysis system, information processing device, and collecting device

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18794065

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2019515746

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18794065

Country of ref document: EP

Kind code of ref document: A1